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 You may also be interested in the [grammar].
26 [book]: book/index.html
27 [standard]: std/index.html
28 [grammar]: grammar.html
32 ## Unicode productions
34 A few productions in Rust's grammar permit Unicode code points outside the
35 ASCII range. We define these productions in terms of character properties
36 specified in the Unicode standard, rather than in terms of ASCII-range code
37 points. The grammar has a [Special Unicode Productions][unicodeproductions]
38 section that lists these productions.
40 [unicodeproductions]: grammar.html#special-unicode-productions
42 ## String table productions
44 Some rules in the grammar — notably [unary
45 operators](#unary-operator-expressions), [binary
46 operators](#binary-operator-expressions), and [keywords][keywords] — are
47 given in a simplified form: as a listing of a table of unquoted, printable
48 whitespace-separated strings. These cases form a subset of the rules regarding
49 the [token](#tokens) rule, and are assumed to be the result of a
50 lexical-analysis phase feeding the parser, driven by a DFA, operating over the
51 disjunction of all such string table entries.
53 [keywords]: grammar.html#keywords
55 When such a string enclosed in double-quotes (`"`) occurs inside the grammar,
56 it is an implicit reference to a single member of such a string table
57 production. See [tokens](#tokens) for more information.
63 Rust input is interpreted as a sequence of Unicode code points encoded in UTF-8.
64 Most Rust grammar rules are defined in terms of printable ASCII-range
65 code points, but a small number are defined in terms of Unicode properties or
66 explicit code point lists. [^inputformat]
68 [^inputformat]: Substitute definitions for the special Unicode productions are
69 provided to the grammar verifier, restricted to ASCII range, when verifying the
70 grammar in this document.
74 An identifier is any nonempty Unicode[^non_ascii_idents] string of the following form:
76 [^non_ascii_idents]: Non-ASCII characters in identifiers are currently feature
77 gated. This is expected to improve soon.
81 * The first character has property `XID_start`
82 * The remaining characters have property `XID_continue`
86 * The first character is `_`
87 * The identifier is more than one character, `_` alone is not an identifier
88 * The remaining characters have property `XID_continue`
90 that does _not_ occur in the set of [keywords][keywords].
92 > **Note**: `XID_start` and `XID_continue` as character properties cover the
93 > character ranges used to form the more familiar C and Java language-family
98 Comments in Rust code follow the general C++ style of line (`//`) and
99 block (`/* ... */`) comment forms. Nested block comments are supported.
101 Line comments beginning with exactly _three_ slashes (`///`), and block
102 comments (`/** ... */`), are interpreted as a special syntax for `doc`
103 [attributes](#attributes). That is, they are equivalent to writing
104 `#[doc="..."]` around the body of the comment, i.e., `/// Foo` turns into
107 Line comments beginning with `//!` and block comments `/*! ... !*/` are
108 doc comments that apply to the parent of the comment, rather than the item
109 that follows. That is, they are equivalent to writing `#![doc="..."]` around
110 the body of the comment. `//!` comments are usually used to document
111 modules that occupy a source file.
113 Non-doc comments are interpreted as a form of whitespace.
117 Whitespace is any non-empty string containing only the following characters:
119 - `U+0020` (space, `' '`)
120 - `U+0009` (tab, `'\t'`)
121 - `U+000A` (LF, `'\n'`)
122 - `U+000D` (CR, `'\r'`)
124 Rust is a "free-form" language, meaning that all forms of whitespace serve only
125 to separate _tokens_ in the grammar, and have no semantic significance.
127 A Rust program has identical meaning if each whitespace element is replaced
128 with any other legal whitespace element, such as a single space character.
132 Tokens are primitive productions in the grammar defined by regular
133 (non-recursive) languages. "Simple" tokens are given in [string table
134 production](#string-table-productions) form, and occur in the rest of the
135 grammar as double-quoted strings. Other tokens have exact rules given.
139 A literal is an expression consisting of a single token, rather than a sequence
140 of tokens, that immediately and directly denotes the value it evaluates to,
141 rather than referring to it by name or some other evaluation rule. A literal is
142 a form of constant expression, so is evaluated (primarily) at compile time.
146 ##### Characters and strings
148 | | Example | `#` sets | Characters | Escapes |
149 |----------------------------------------------|-----------------|------------|-------------|---------------------|
150 | [Character](#character-literals) | `'H'` | `N/A` | All Unicode | [Quote](#quote-escapes) & [Byte](#byte-escapes) & [Unicode](#unicode-escapes) |
151 | [String](#string-literals) | `"hello"` | `N/A` | All Unicode | [Quote](#quote-escapes) & [Byte](#byte-escapes) & [Unicode](#unicode-escapes) |
152 | [Raw](#raw-string-literals) | `r#"hello"#` | `0...` | All Unicode | `N/A` |
153 | [Byte](#byte-literals) | `b'H'` | `N/A` | All ASCII | [Quote](#quote-escapes) & [Byte](#byte-escapes) |
154 | [Byte string](#byte-string-literals) | `b"hello"` | `N/A` | All ASCII | [Quote](#quote-escapes) & [Byte](#byte-escapes) |
155 | [Raw byte string](#raw-byte-string-literals) | `br#"hello"#` | `0...` | All ASCII | `N/A` |
161 | `\x7F` | 8-bit character code (exactly 2 digits) |
163 | `\r` | Carriage return |
168 ##### Unicode escapes
171 | `\u{7FFF}` | 24-bit Unicode character code (up to 6 digits) |
176 | `\'` | Single quote |
177 | `\"` | Double quote |
181 | [Number literals](#number-literals)`*` | Example | Exponentiation | Suffixes |
182 |----------------------------------------|---------|----------------|----------|
183 | Decimal integer | `98_222` | `N/A` | Integer suffixes |
184 | Hex integer | `0xff` | `N/A` | Integer suffixes |
185 | Octal integer | `0o77` | `N/A` | Integer suffixes |
186 | Binary integer | `0b1111_0000` | `N/A` | Integer suffixes |
187 | Floating-point | `123.0E+77` | `Optional` | Floating-point suffixes |
189 `*` All number literals allow `_` as a visual separator: `1_234.0E+18f64`
192 | Integer | Floating-point |
193 |---------|----------------|
194 | `u8`, `i8`, `u16`, `i16`, `u32`, `i32`, `u64`, `i64`, `isize`, `usize` | `f32`, `f64` |
196 #### Character and string literals
198 ##### Character literals
200 A _character literal_ is a single Unicode character enclosed within two
201 `U+0027` (single-quote) characters, with the exception of `U+0027` itself,
202 which must be _escaped_ by a preceding `U+005C` character (`\`).
204 ##### String literals
206 A _string literal_ is a sequence of any Unicode characters enclosed within two
207 `U+0022` (double-quote) characters, with the exception of `U+0022` itself,
208 which must be _escaped_ by a preceding `U+005C` character (`\`).
210 Line-break characters are allowed in string literals. Normally they represent
211 themselves (i.e. no translation), but as a special exception, when a `U+005C`
212 character (`\`) occurs immediately before the newline, the `U+005C` character,
213 the newline, and all whitespace at the beginning of the next line are ignored.
214 Thus `a` and `b` are equal:
224 ##### Character escapes
226 Some additional _escapes_ are available in either character or non-raw string
227 literals. An escape starts with a `U+005C` (`\`) and continues with one of the
230 * An _8-bit code point escape_ starts with `U+0078` (`x`) and is
231 followed by exactly two _hex digits_. It denotes the Unicode code point
232 equal to the provided hex value.
233 * A _24-bit code point escape_ starts with `U+0075` (`u`) and is followed
234 by up to six _hex digits_ surrounded by braces `U+007B` (`{`) and `U+007D`
235 (`}`). It denotes the Unicode code point equal to the provided hex value.
236 * A _whitespace escape_ is one of the characters `U+006E` (`n`), `U+0072`
237 (`r`), or `U+0074` (`t`), denoting the Unicode values `U+000A` (LF),
238 `U+000D` (CR) or `U+0009` (HT) respectively.
239 * The _backslash escape_ is the character `U+005C` (`\`) which must be
240 escaped in order to denote *itself*.
242 ##### Raw string literals
244 Raw string literals do not process any escapes. They start with the character
245 `U+0072` (`r`), followed by zero or more of the character `U+0023` (`#`) and a
246 `U+0022` (double-quote) character. The _raw string body_ can contain any sequence
247 of Unicode characters and is terminated only by another `U+0022` (double-quote)
248 character, followed by the same number of `U+0023` (`#`) characters that preceded
249 the opening `U+0022` (double-quote) character.
251 All Unicode characters contained in the raw string body represent themselves,
252 the characters `U+0022` (double-quote) (except when followed by at least as
253 many `U+0023` (`#`) characters as were used to start the raw string literal) or
254 `U+005C` (`\`) do not have any special meaning.
256 Examples for string literals:
259 "foo"; r"foo"; // foo
260 "\"foo\""; r#""foo""#; // "foo"
263 r##"foo #"# bar"##; // foo #"# bar
265 "\x52"; "R"; r"R"; // R
266 "\\x52"; r"\x52"; // \x52
269 #### Byte and byte string literals
273 A _byte literal_ is a single ASCII character (in the `U+0000` to `U+007F`
274 range) or a single _escape_ preceded by the characters `U+0062` (`b`) and
275 `U+0027` (single-quote), and followed by the character `U+0027`. If the character
276 `U+0027` is present within the literal, it must be _escaped_ by a preceding
277 `U+005C` (`\`) character. It is equivalent to a `u8` unsigned 8-bit integer
280 ##### Byte string literals
282 A non-raw _byte string literal_ is a sequence of ASCII characters and _escapes_,
283 preceded by the characters `U+0062` (`b`) and `U+0022` (double-quote), and
284 followed by the character `U+0022`. If the character `U+0022` is present within
285 the literal, it must be _escaped_ by a preceding `U+005C` (`\`) character.
286 Alternatively, a byte string literal can be a _raw byte string literal_, defined
287 below. A byte string literal of length `n` is equivalent to a `&'static [u8; n]` borrowed fixed-sized array
288 of unsigned 8-bit integers.
290 Some additional _escapes_ are available in either byte or non-raw byte string
291 literals. An escape starts with a `U+005C` (`\`) and continues with one of the
294 * A _byte escape_ escape starts with `U+0078` (`x`) and is
295 followed by exactly two _hex digits_. It denotes the byte
296 equal to the provided hex value.
297 * A _whitespace escape_ is one of the characters `U+006E` (`n`), `U+0072`
298 (`r`), or `U+0074` (`t`), denoting the bytes values `0x0A` (ASCII LF),
299 `0x0D` (ASCII CR) or `0x09` (ASCII HT) respectively.
300 * The _backslash escape_ is the character `U+005C` (`\`) which must be
301 escaped in order to denote its ASCII encoding `0x5C`.
303 ##### Raw byte string literals
305 Raw byte string literals do not process any escapes. They start with the
306 character `U+0062` (`b`), followed by `U+0072` (`r`), followed by zero or more
307 of the character `U+0023` (`#`), and a `U+0022` (double-quote) character. The
308 _raw string body_ can contain any sequence of ASCII characters and is terminated
309 only by another `U+0022` (double-quote) character, followed by the same number of
310 `U+0023` (`#`) characters that preceded the opening `U+0022` (double-quote)
311 character. A raw byte string literal can not contain any non-ASCII byte.
313 All characters contained in the raw string body represent their ASCII encoding,
314 the characters `U+0022` (double-quote) (except when followed by at least as
315 many `U+0023` (`#`) characters as were used to start the raw string literal) or
316 `U+005C` (`\`) do not have any special meaning.
318 Examples for byte string literals:
321 b"foo"; br"foo"; // foo
322 b"\"foo\""; br#""foo""#; // "foo"
325 br##"foo #"# bar"##; // foo #"# bar
327 b"\x52"; b"R"; br"R"; // R
328 b"\\x52"; br"\x52"; // \x52
333 A _number literal_ is either an _integer literal_ or a _floating-point
334 literal_. The grammar for recognizing the two kinds of literals is mixed.
336 ##### Integer literals
338 An _integer literal_ has one of four forms:
340 * A _decimal literal_ starts with a *decimal digit* and continues with any
341 mixture of *decimal digits* and _underscores_.
342 * A _hex literal_ starts with the character sequence `U+0030` `U+0078`
343 (`0x`) and continues as any mixture of hex digits and underscores.
344 * An _octal literal_ starts with the character sequence `U+0030` `U+006F`
345 (`0o`) and continues as any mixture of octal digits and underscores.
346 * A _binary literal_ starts with the character sequence `U+0030` `U+0062`
347 (`0b`) and continues as any mixture of binary digits and underscores.
349 Like any literal, an integer literal may be followed (immediately,
350 without any spaces) by an _integer suffix_, which forcibly sets the
351 type of the literal. The integer suffix must be the name of one of the
352 integral types: `u8`, `i8`, `u16`, `i16`, `u32`, `i32`, `u64`, `i64`,
355 The type of an _unsuffixed_ integer literal is determined by type inference:
357 * If an integer type can be _uniquely_ determined from the surrounding
358 program context, the unsuffixed integer literal has that type.
360 * If the program context under-constrains the type, it defaults to the
361 signed 32-bit integer `i32`.
363 * If the program context over-constrains the type, it is considered a
366 Examples of integer literals of various forms:
373 0o70_i16; // type i16
374 0b1111_1111_1001_0000_i32; // type i32
375 0usize; // type usize
378 ##### Floating-point literals
380 A _floating-point literal_ has one of two forms:
382 * A _decimal literal_ followed by a period character `U+002E` (`.`). This is
383 optionally followed by another decimal literal, with an optional _exponent_.
384 * A single _decimal literal_ followed by an _exponent_.
386 Like integer literals, a floating-point literal may be followed by a
387 suffix, so long as the pre-suffix part does not end with `U+002E` (`.`).
388 The suffix forcibly sets the type of the literal. There are two valid
389 _floating-point suffixes_, `f32` and `f64` (the 32-bit and 64-bit floating point
390 types), which explicitly determine the type of the literal.
392 The type of an _unsuffixed_ floating-point literal is determined by
395 * If a floating-point type can be _uniquely_ determined from the
396 surrounding program context, the unsuffixed floating-point literal
399 * If the program context under-constrains the type, it defaults to `f64`.
401 * If the program context over-constrains the type, it is considered a
404 Examples of floating-point literals of various forms:
407 123.0f64; // type f64
410 12E+99_f64; // type f64
411 let x: f64 = 2.; // type f64
414 This last example is different because it is not possible to use the suffix
415 syntax with a floating point literal ending in a period. `2.f64` would attempt
416 to call a method named `f64` on `2`.
418 The representation semantics of floating-point numbers are described in
419 ["Machine Types"](#machine-types).
421 #### Boolean literals
423 The two values of the boolean type are written `true` and `false`.
427 Symbols are a general class of printable [tokens](#tokens) that play structural
428 roles in a variety of grammar productions. They are a
429 set of remaining miscellaneous printable tokens that do not
430 otherwise appear as [unary operators](#unary-operator-expressions), [binary
431 operators](#binary-operator-expressions), or [keywords][keywords].
432 They are catalogued in [the Symbols section][symbols] of the Grammar document.
434 [symbols]: grammar.html#symbols
439 A _path_ is a sequence of one or more path components _logically_ separated by
440 a namespace qualifier (`::`). If a path consists of only one component, it may
441 refer to either an [item](#items) or a [variable](#variables) in a local control
442 scope. If a path has multiple components, it refers to an item.
444 Every item has a _canonical path_ within its crate, but the path naming an item
445 is only meaningful within a given crate. There is no global namespace across
446 crates; an item's canonical path merely identifies it within the crate.
448 Two examples of simple paths consisting of only identifier components:
455 Path components are usually [identifiers](#identifiers), but they may
456 also include angle-bracket-enclosed lists of type arguments. In
457 [expression](#expressions) context, the type argument list is given
458 after a `::` namespace qualifier in order to disambiguate it from a
459 relational expression involving the less-than symbol (`<`). In type
460 expression context, the final namespace qualifier is omitted.
462 Two examples of paths with type arguments:
465 # struct HashMap<K, V>(K,V);
467 # fn id<T>(t: T) -> T { t }
468 type T = HashMap<i32,String>; // Type arguments used in a type expression
469 let x = id::<i32>(10); // Type arguments used in a call expression
473 Paths can be denoted with various leading qualifiers to change the meaning of
476 * Paths starting with `::` are considered to be global paths where the
477 components of the path start being resolved from the crate root. Each
478 identifier in the path must resolve to an item.
486 ::a::foo(); // call a's foo function
492 * Paths starting with the keyword `super` begin resolution relative to the
493 parent module. Each further identifier must resolve to an item.
501 super::a::foo(); // call a's foo function
507 * Paths starting with the keyword `self` begin resolution relative to the
508 current module. Each further identifier must resolve to an item.
520 A number of minor features of Rust are not central enough to have their own
521 syntax, and yet are not implementable as functions. Instead, they are given
522 names, and invoked through a consistent syntax: `some_extension!(...)`.
524 Users of `rustc` can define new syntax extensions in two ways:
526 * [Compiler plugins][plugin] can include arbitrary Rust code that
527 manipulates syntax trees at compile time. Note that the interface
528 for compiler plugins is considered highly unstable.
530 * [Macros](book/macros.html) define new syntax in a higher-level,
535 `macro_rules` allows users to define syntax extension in a declarative way. We
536 call such extensions "macros by example" or simply "macros" — to be distinguished
537 from the "procedural macros" defined in [compiler plugins][plugin].
539 Currently, macros can expand to expressions, statements, items, or patterns.
541 (A `sep_token` is any token other than `*` and `+`. A `non_special_token` is
542 any token other than a delimiter or `$`.)
544 The macro expander looks up macro invocations by name, and tries each macro
545 rule in turn. It transcribes the first successful match. Matching and
546 transcription are closely related to each other, and we will describe them
551 The macro expander matches and transcribes every token that does not begin with
552 a `$` literally, including delimiters. For parsing reasons, delimiters must be
553 balanced, but they are otherwise not special.
555 In the matcher, `$` _name_ `:` _designator_ matches the nonterminal in the Rust
556 syntax named by _designator_. Valid designators are:
558 * `item`: an [item](#items)
559 * `block`: a [block](#block-expressions)
560 * `stmt`: a [statement](#statements)
561 * `pat`: a [pattern](#match-expressions)
562 * `expr`: an [expression](#expressions)
563 * `ty`: a [type](#types)
564 * `ident`: an [identifier](#identifiers)
565 * `path`: a [path](#paths)
566 * `tt`: either side of the `=>` in macro rules
567 * `meta`: the contents of an [attribute](#attributes)
569 In the transcriber, the
570 designator is already known, and so only the name of a matched nonterminal comes
571 after the dollar sign.
573 In both the matcher and transcriber, the Kleene star-like operator indicates
574 repetition. The Kleene star operator consists of `$` and parentheses, optionally
575 followed by a separator token, followed by `*` or `+`. `*` means zero or more
576 repetitions, `+` means at least one repetition. The parentheses are not matched or
577 transcribed. On the matcher side, a name is bound to _all_ of the names it
578 matches, in a structure that mimics the structure of the repetition encountered
579 on a successful match. The job of the transcriber is to sort that structure
582 The rules for transcription of these repetitions are called "Macro By Example".
583 Essentially, one "layer" of repetition is discharged at a time, and all of them
584 must be discharged by the time a name is transcribed. Therefore, `( $( $i:ident
585 ),* ) => ( $i )` is an invalid macro, but `( $( $i:ident ),* ) => ( $( $i:ident
586 ),* )` is acceptable (if trivial).
588 When Macro By Example encounters a repetition, it examines all of the `$`
589 _name_ s that occur in its body. At the "current layer", they all must repeat
590 the same number of times, so ` ( $( $i:ident ),* ; $( $j:ident ),* ) => ( $(
591 ($i,$j) ),* )` is valid if given the argument `(a,b,c ; d,e,f)`, but not
592 `(a,b,c ; d,e)`. The repetition walks through the choices at that layer in
593 lockstep, so the former input transcribes to `(a,d), (b,e), (c,f)`.
595 Nested repetitions are allowed.
597 ### Parsing limitations
599 The parser used by the macro system is reasonably powerful, but the parsing of
600 Rust syntax is restricted in two ways:
602 1. Macro definitions are required to include suitable separators after parsing
603 expressions and other bits of the Rust grammar. This implies that
604 a macro definition like `$i:expr [ , ]` is not legal, because `[` could be part
605 of an expression. A macro definition like `$i:expr,` or `$i:expr;` would be legal,
606 however, because `,` and `;` are legal separators. See [RFC 550] for more information.
607 2. The parser must have eliminated all ambiguity by the time it reaches a `$`
608 _name_ `:` _designator_. This requirement most often affects name-designator
609 pairs when they occur at the beginning of, or immediately after, a `$(...)*`;
610 requiring a distinctive token in front can solve the problem.
612 [RFC 550]: https://github.com/rust-lang/rfcs/blob/master/text/0550-macro-future-proofing.md
614 # Crates and source files
616 Although Rust, like any other language, can be implemented by an interpreter as
617 well as a compiler, the only existing implementation is a compiler,
619 always been designed to be compiled. For these reasons, this section assumes a
622 Rust's semantics obey a *phase distinction* between compile-time and
623 run-time.[^phase-distinction] Semantic rules that have a *static
624 interpretation* govern the success or failure of compilation, while
626 that have a *dynamic interpretation* govern the behavior of the program at
629 [^phase-distinction]: This distinction would also exist in an interpreter.
630 Static checks like syntactic analysis, type checking, and lints should
631 happen before the program is executed regardless of when it is executed.
633 The compilation model centers on artifacts called _crates_. Each compilation
634 processes a single crate in source form, and if successful, produces a single
635 crate in binary form: either an executable or some sort of
636 library.[^cratesourcefile]
638 [^cratesourcefile]: A crate is somewhat analogous to an *assembly* in the
639 ECMA-335 CLI model, a *library* in the SML/NJ Compilation Manager, a *unit*
640 in the Owens and Flatt module system, or a *configuration* in Mesa.
642 A _crate_ is a unit of compilation and linking, as well as versioning,
643 distribution and runtime loading. A crate contains a _tree_ of nested
644 [module](#modules) scopes. The top level of this tree is a module that is
645 anonymous (from the point of view of paths within the module) and any item
646 within a crate has a canonical [module path](#paths) denoting its location
647 within the crate's module tree.
649 The Rust compiler is always invoked with a single source file as input, and
650 always produces a single output crate. The processing of that source file may
651 result in other source files being loaded as modules. Source files have the
654 A Rust source file describes a module, the name and location of which —
655 in the module tree of the current crate — are defined from outside the
656 source file: either by an explicit `mod_item` in a referencing source file, or
657 by the name of the crate itself. Every source file is a module, but not every
658 module needs its own source file: [module definitions](#modules) can be nested
661 Each source file contains a sequence of zero or more `item` definitions, and
662 may optionally begin with any number of [attributes](#items-and-attributes)
663 that apply to the containing module, most of which influence the behavior of
664 the compiler. The anonymous crate module can have additional attributes that
665 apply to the crate as a whole.
668 // Specify the crate name.
669 #![crate_name = "projx"]
671 // Specify the type of output artifact.
672 #![crate_type = "lib"]
674 // Turn on a warning.
675 // This can be done in any module, not just the anonymous crate module.
676 #![warn(non_camel_case_types)]
679 A crate that contains a `main` function can be compiled to an executable. If a
680 `main` function is present, its return type must be `()`
681 ("[unit](#tuple-types)") and it must take no arguments.
683 # Items and attributes
685 Crates contain [items](#items), each of which may have some number of
686 [attributes](#attributes) attached to it.
690 An _item_ is a component of a crate. Items are organized within a crate by a
691 nested set of [modules](#modules). Every crate has a single "outermost"
692 anonymous module; all further items within the crate have [paths](#paths)
693 within the module tree of the crate.
695 Items are entirely determined at compile-time, generally remain fixed during
696 execution, and may reside in read-only memory.
698 There are several kinds of item:
700 * [`extern crate` declarations](#extern-crate-declarations)
701 * [`use` declarations](#use-declarations)
702 * [modules](#modules)
703 * [functions](#functions)
704 * [type definitions](grammar.html#type-definitions)
705 * [structs](#structs)
706 * [enumerations](#enumerations)
707 * [constant items](#constant-items)
708 * [static items](#static-items)
710 * [implementations](#implementations)
712 Some items form an implicit scope for the declaration of sub-items. In other
713 words, within a function or module, declarations of items can (in many cases)
714 be mixed with the statements, control blocks, and similar artifacts that
715 otherwise compose the item body. The meaning of these scoped items is the same
716 as if the item was declared outside the scope — it is still a static item
717 — except that the item's *path name* within the module namespace is
718 qualified by the name of the enclosing item, or is private to the enclosing
719 item (in the case of functions). The grammar specifies the exact locations in
720 which sub-item declarations may appear.
724 All items except modules, constants and statics may be *parameterized* by type.
725 Type parameters are given as a comma-separated list of identifiers enclosed in
726 angle brackets (`<...>`), after the name of the item and before its definition.
727 The type parameters of an item are considered "part of the name", not part of
728 the type of the item. A referencing [path](#paths) must (in principle) provide
729 type arguments as a list of comma-separated types enclosed within angle
730 brackets, in order to refer to the type-parameterized item. In practice, the
731 type-inference system can usually infer such argument types from context. There
732 are no general type-parametric types, only type-parametric items. That is, Rust
733 has no notion of type abstraction: there are no higher-ranked (or "forall") types
734 abstracted over other types, though higher-ranked types do exist for lifetimes.
738 A module is a container for zero or more [items](#items).
740 A _module item_ is a module, surrounded in braces, named, and prefixed with the
741 keyword `mod`. A module item introduces a new, named module into the tree of
742 modules making up a crate. Modules can nest arbitrarily.
744 An example of a module:
748 type Complex = (f64, f64);
749 fn sin(f: f64) -> f64 {
753 fn cos(f: f64) -> f64 {
757 fn tan(f: f64) -> f64 {
764 Modules and types share the same namespace. Declaring a named type with
765 the same name as a module in scope is forbidden: that is, a type definition,
766 trait, struct, enumeration, or type parameter can't shadow the name of a module
767 in scope, or vice versa.
769 A module without a body is loaded from an external file, by default with the
770 same name as the module, plus the `.rs` extension. When a nested submodule is
771 loaded from an external file, it is loaded from a subdirectory path that
772 mirrors the module hierarchy.
775 // Load the `vec` module from `vec.rs`
779 // Load the `local_data` module from `thread/local_data.rs`
780 // or `thread/local_data/mod.rs`.
785 The directories and files used for loading external file modules can be
786 influenced with the `path` attribute.
789 #[path = "thread_files"]
791 // Load the `local_data` module from `thread_files/tls.rs`
797 #### Extern crate declarations
799 An _`extern crate` declaration_ specifies a dependency on an external crate.
800 The external crate is then bound into the declaring scope as the `ident`
801 provided in the `extern_crate_decl`.
803 The external crate is resolved to a specific `soname` at compile time, and a
804 runtime linkage requirement to that `soname` is passed to the linker for
805 loading at runtime. The `soname` is resolved at compile time by scanning the
806 compiler's library path and matching the optional `crateid` provided against
807 the `crateid` attributes that were declared on the external crate when it was
808 compiled. If no `crateid` is provided, a default `name` attribute is assumed,
809 equal to the `ident` given in the `extern_crate_decl`.
811 Three examples of `extern crate` declarations:
816 extern crate std; // equivalent to: extern crate std as std;
818 extern crate std as ruststd; // linking to 'std' under another name
821 #### Use declarations
823 A _use declaration_ creates one or more local name bindings synonymous with
824 some other [path](#paths). Usually a `use` declaration is used to shorten the
825 path required to refer to a module item. These declarations may appear at the
826 top of [modules](#modules) and [blocks](grammar.html#block-expressions).
828 > **Note**: Unlike in many languages,
829 > `use` declarations in Rust do *not* declare linkage dependency with external crates.
830 > Rather, [`extern crate` declarations](#extern-crate-declarations) declare linkage dependencies.
832 Use declarations support a number of convenient shortcuts:
834 * Rebinding the target name as a new local name, using the syntax `use p::q::r as x;`
835 * Simultaneously binding a list of paths differing only in their final element,
836 using the glob-like brace syntax `use a::b::{c,d,e,f};`
837 * Binding all paths matching a given prefix, using the asterisk wildcard syntax
839 * Simultaneously binding a list of paths differing only in their final element
840 and their immediate parent module, using the `self` keyword, such as
841 `use a::b::{self, c, d};`
843 An example of `use` declarations:
846 use std::option::Option::{Some, None};
847 use std::collections::hash_map::{self, HashMap};
850 fn bar(map1: HashMap<String, usize>, map2: hash_map::HashMap<String, usize>){}
853 // Equivalent to 'foo(vec![std::option::Option::Some(1.0f64),
854 // std::option::Option::None]);'
855 foo(vec![Some(1.0f64), None]);
857 // Both `hash_map` and `HashMap` are in scope.
858 let map1 = HashMap::new();
859 let map2 = hash_map::HashMap::new();
864 Like items, `use` declarations are private to the containing module, by
865 default. Also like items, a `use` declaration can be public, if qualified by
866 the `pub` keyword. Such a `use` declaration serves to _re-export_ a name. A
867 public `use` declaration can therefore _redirect_ some public name to a
868 different target definition: even a definition with a private canonical path,
869 inside a different module. If a sequence of such redirections form a cycle or
870 cannot be resolved unambiguously, they represent a compile-time error.
872 An example of re-exporting:
877 pub use quux::foo::{bar, baz};
886 In this example, the module `quux` re-exports two public names defined in
889 Also note that the paths contained in `use` items are relative to the crate
890 root. So, in the previous example, the `use` refers to `quux::foo::{bar,
891 baz}`, and not simply to `foo::{bar, baz}`. This also means that top-level
892 module declarations should be at the crate root if direct usage of the declared
893 modules within `use` items is desired. It is also possible to use `self` and
894 `super` at the beginning of a `use` item to refer to the current and direct
895 parent modules respectively. All rules regarding accessing declared modules in
896 `use` declarations apply to both module declarations and `extern crate`
899 An example of what will and will not work for `use` items:
902 # #![allow(unused_imports)]
903 use foo::baz::foobaz; // good: foo is at the root of the crate
911 use foo::example::iter; // good: foo is at crate root
912 // use example::iter; // bad: example is not at the crate root
913 use self::baz::foobaz; // good: self refers to module 'foo'
914 use foo::bar::foobar; // good: foo is at crate root
921 use super::bar::foobar; // good: super refers to module 'foo'
931 A _function item_ defines a sequence of [statements](#statements) and a
932 final [expression](#expressions), along with a name and a set of
933 parameters. Other than a name, all these are optional.
934 Functions are declared with the keyword `fn`. Functions may declare a
935 set of *input* [*variables*](#variables) as parameters, through which the caller
936 passes arguments into the function, and the *output* [*type*](#types)
937 of the value the function will return to its caller on completion.
939 A function may also be copied into a first-class *value*, in which case the
940 value has the corresponding [*function type*](#function-types), and can be used
941 otherwise exactly as a function item (with a minor additional cost of calling
942 the function indirectly).
944 Every control path in a function logically ends with a `return` expression or a
945 diverging expression. If the outermost block of a function has a
946 value-producing expression in its final-expression position, that expression is
947 interpreted as an implicit `return` expression applied to the final-expression.
949 An example of a function:
952 fn add(x: i32, y: i32) -> i32 {
957 As with `let` bindings, function arguments are irrefutable patterns, so any
958 pattern that is valid in a let binding is also valid as an argument.
961 fn first((value, _): (i32, i32)) -> i32 { value }
965 #### Generic functions
967 A _generic function_ allows one or more _parameterized types_ to appear in its
968 signature. Each type parameter must be explicitly declared, in an
969 angle-bracket-enclosed, comma-separated list following the function name.
972 // foo is generic over A and B
974 fn foo<A, B>(x: A, y: B) {
977 Inside the function signature and body, the name of the type parameter can be
978 used as a type name. [Trait](#traits) bounds can be specified for type parameters
979 to allow methods with that trait to be called on values of that type. This is
980 specified using the `where` syntax:
983 fn foo<T>(x: T) where T: Debug {
986 When a generic function is referenced, its type is instantiated based on the
987 context of the reference. For example, calling the `foo` function here:
992 fn foo<T>(x: &[T]) where T: Debug {
1000 will instantiate type parameter `T` with `i32`.
1002 The type parameters can also be explicitly supplied in a trailing
1003 [path](#paths) component after the function name. This might be necessary if
1004 there is not sufficient context to determine the type parameters. For example,
1005 `mem::size_of::<u32>() == 4`.
1007 #### Diverging functions
1009 A special kind of function can be declared with a `!` character where the
1010 output type would normally be. For example:
1013 fn my_err(s: &str) -> ! {
1019 We call such functions "diverging" because they never return a value to the
1020 caller. Every control path in a diverging function must end with a `panic!()` or
1021 a call to another diverging function on every control path. The `!` annotation
1022 does *not* denote a type.
1024 It might be necessary to declare a diverging function because as mentioned
1025 previously, the typechecker checks that every control path in a function ends
1026 with a [`return`](#return-expressions) or diverging expression. So, if `my_err`
1027 were declared without the `!` annotation, the following code would not
1031 # fn my_err(s: &str) -> ! { panic!() }
1033 fn f(i: i32) -> i32 {
1038 my_err("Bad number!");
1043 This will not compile without the `!` annotation on `my_err`, since the `else`
1044 branch of the conditional in `f` does not return an `i32`, as required by the
1045 signature of `f`. Adding the `!` annotation to `my_err` informs the
1046 typechecker that, should control ever enter `my_err`, no further type judgments
1047 about `f` need to hold, since control will never resume in any context that
1048 relies on those judgments. Thus the return type on `f` only needs to reflect
1049 the `if` branch of the conditional.
1051 #### Extern functions
1053 Extern functions are part of Rust's foreign function interface, providing the
1054 opposite functionality to [external blocks](#external-blocks). Whereas
1055 external blocks allow Rust code to call foreign code, extern functions with
1056 bodies defined in Rust code _can be called by foreign code_. They are defined
1057 in the same way as any other Rust function, except that they have the `extern`
1061 // Declares an extern fn, the ABI defaults to "C"
1062 extern fn new_i32() -> i32 { 0 }
1064 // Declares an extern fn with "stdcall" ABI
1065 extern "stdcall" fn new_i32_stdcall() -> i32 { 0 }
1068 Unlike normal functions, extern fns have type `extern "ABI" fn()`. This is the
1069 same type as the functions declared in an extern block.
1072 # extern fn new_i32() -> i32 { 0 }
1073 let fptr: extern "C" fn() -> i32 = new_i32;
1076 Extern functions may be called directly from Rust code as Rust uses large,
1077 contiguous stack segments like C.
1081 A _type alias_ defines a new name for an existing [type](#types). Type
1082 aliases are declared with the keyword `type`. Every value has a single,
1083 specific type, but may implement several different traits, or be compatible with
1084 several different type constraints.
1086 For example, the following defines the type `Point` as a synonym for the type
1087 `(u8, u8)`, the type of pairs of unsigned 8 bit integers:
1090 type Point = (u8, u8);
1091 let p: Point = (41, 68);
1096 A _struct_ is a nominal [struct type](#struct-types) defined with the
1099 An example of a `struct` item and its use:
1102 struct Point {x: i32, y: i32}
1103 let p = Point {x: 10, y: 11};
1107 A _tuple struct_ is a nominal [tuple type](#tuple-types), also defined with
1108 the keyword `struct`. For example:
1111 struct Point(i32, i32);
1112 let p = Point(10, 11);
1113 let px: i32 = match p { Point(x, _) => x };
1116 A _unit-like struct_ is a struct without any fields, defined by leaving off
1117 the list of fields entirely. Such a struct implicitly defines a constant of
1118 its type with the same name. For example:
1121 # #![feature(braced_empty_structs)]
1123 let c = [Cookie, Cookie {}, Cookie, Cookie {}];
1129 # #![feature(braced_empty_structs)]
1131 const Cookie: Cookie = Cookie {};
1132 let c = [Cookie, Cookie {}, Cookie, Cookie {}];
1135 The precise memory layout of a struct is not specified. One can specify a
1136 particular layout using the [`repr` attribute](#ffi-attributes).
1140 An _enumeration_ is a simultaneous definition of a nominal [enumerated
1141 type](#enumerated-types) as well as a set of *constructors*, that can be used
1142 to create or pattern-match values of the corresponding enumerated type.
1144 Enumerations are declared with the keyword `enum`.
1146 An example of an `enum` item and its use:
1154 let mut a: Animal = Animal::Dog;
1158 Enumeration constructors can have either named or unnamed fields:
1163 Cat { name: String, weight: f64 }
1166 let mut a: Animal = Animal::Dog("Cocoa".to_string(), 37.2);
1167 a = Animal::Cat { name: "Spotty".to_string(), weight: 2.7 };
1170 In this example, `Cat` is a _struct-like enum variant_,
1171 whereas `Dog` is simply called an enum variant.
1173 Enums have a discriminant. You can assign them explicitly:
1181 If a discriminant isn't assigned, they start at zero, and add one for each
1184 You can cast an enum to get this value:
1187 # enum Foo { Bar = 123 }
1188 let x = Foo::Bar as u32; // x is now 123u32
1191 This only works as long as none of the variants have data attached. If
1192 it were `Bar(i32)`, this is disallowed.
1196 A *constant item* is a named _constant value_ which is not associated with a
1197 specific memory location in the program. Constants are essentially inlined
1198 wherever they are used, meaning that they are copied directly into the relevant
1199 context when used. References to the same constant are not necessarily
1200 guaranteed to refer to the same memory address.
1202 Constant values must not have destructors, and otherwise permit most forms of
1203 data. Constants may refer to the address of other constants, in which case the
1204 address will have the `static` lifetime. The compiler is, however, still at
1205 liberty to translate the constant many times, so the address referred to may not
1208 Constants must be explicitly typed. The type may be `bool`, `char`, a number, or
1209 a type derived from those primitive types. The derived types are references with
1210 the `static` lifetime, fixed-size arrays, tuples, enum variants, and structs.
1213 const BIT1: u32 = 1 << 0;
1214 const BIT2: u32 = 1 << 1;
1216 const BITS: [u32; 2] = [BIT1, BIT2];
1217 const STRING: &'static str = "bitstring";
1219 struct BitsNStrings<'a> {
1224 const BITS_N_STRINGS: BitsNStrings<'static> = BitsNStrings {
1232 A *static item* is similar to a *constant*, except that it represents a precise
1233 memory location in the program. A static is never "inlined" at the usage site,
1234 and all references to it refer to the same memory location. Static items have
1235 the `static` lifetime, which outlives all other lifetimes in a Rust program.
1236 Static items may be placed in read-only memory if they do not contain any
1237 interior mutability.
1239 Statics may contain interior mutability through the `UnsafeCell` language item.
1240 All access to a static is safe, but there are a number of restrictions on
1243 * Statics may not contain any destructors.
1244 * The types of static values must ascribe to `Sync` to allow thread-safe access.
1245 * Statics may not refer to other statics by value, only by reference.
1246 * Constants cannot refer to statics.
1248 Constants should in general be preferred over statics, unless large amounts of
1249 data are being stored, or single-address and mutability properties are required.
1251 #### Mutable statics
1253 If a static item is declared with the `mut` keyword, then it is allowed to
1254 be modified by the program. One of Rust's goals is to make concurrency bugs
1255 hard to run into, and this is obviously a very large source of race conditions
1256 or other bugs. For this reason, an `unsafe` block is required when either
1257 reading or writing a mutable static variable. Care should be taken to ensure
1258 that modifications to a mutable static are safe with respect to other threads
1259 running in the same process.
1261 Mutable statics are still very useful, however. They can be used with C
1262 libraries and can also be bound from C libraries (in an `extern` block).
1265 # fn atomic_add(_: &mut u32, _: u32) -> u32 { 2 }
1267 static mut LEVELS: u32 = 0;
1269 // This violates the idea of no shared state, and this doesn't internally
1270 // protect against races, so this function is `unsafe`
1271 unsafe fn bump_levels_unsafe1() -> u32 {
1277 // Assuming that we have an atomic_add function which returns the old value,
1278 // this function is "safe" but the meaning of the return value may not be what
1279 // callers expect, so it's still marked as `unsafe`
1280 unsafe fn bump_levels_unsafe2() -> u32 {
1281 return atomic_add(&mut LEVELS, 1);
1285 Mutable statics have the same restrictions as normal statics, except that the
1286 type of the value is not required to ascribe to `Sync`.
1290 A _trait_ describes an abstract interface that types can
1291 implement. This interface consists of associated items, which come in
1298 Associated functions whose first parameter is named `self` are called
1299 methods and may be invoked using `.` notation (e.g., `x.foo()`).
1301 All traits define an implicit type parameter `Self` that refers to
1302 "the type that is implementing this interface". Traits may also
1303 contain additional type parameters. These type parameters (including
1304 `Self`) may be constrained by other traits and so forth as usual.
1306 Trait bounds on `Self` are considered "supertraits". These are
1307 required to be acyclic. Supertraits are somewhat different from other
1308 constraints in that they affect what methods are available in the
1309 vtable when the trait is used as a [trait object](#trait-objects).
1311 Traits are implemented for specific types through separate
1312 [implementations](#implementations).
1314 Consider the following trait:
1317 # type Surface = i32;
1318 # type BoundingBox = i32;
1320 fn draw(&self, Surface);
1321 fn bounding_box(&self) -> BoundingBox;
1325 This defines a trait with two methods. All values that have
1326 [implementations](#implementations) of this trait in scope can have their
1327 `draw` and `bounding_box` methods called, using `value.bounding_box()`
1328 [syntax](#method-call-expressions).
1330 Traits can include default implementations of methods, as in:
1335 fn baz(&self) { println!("We called baz."); }
1339 Here the `baz` method has a default implementation, so types that implement
1340 `Foo` need only implement `bar`. It is also possible for implementing types
1341 to override a method that has a default implementation.
1343 Type parameters can be specified for a trait to make it generic. These appear
1344 after the trait name, using the same syntax used in [generic
1345 functions](#generic-functions).
1349 fn len(&self) -> u32;
1350 fn elt_at(&self, n: u32) -> T;
1351 fn iter<F>(&self, F) where F: Fn(T);
1355 It is also possible to define associated types for a trait. Consider the
1356 following example of a `Container` trait. Notice how the type is available
1357 for use in the method signatures:
1363 fn insert(&mut self, Self::E);
1367 In order for a type to implement this trait, it must not only provide
1368 implementations for every method, but it must specify the type `E`. Here's
1369 an implementation of `Container` for the standard library type `Vec`:
1374 # fn empty() -> Self;
1375 # fn insert(&mut self, Self::E);
1377 impl<T> Container for Vec<T> {
1379 fn empty() -> Vec<T> { Vec::new() }
1380 fn insert(&mut self, x: T) { self.push(x); }
1384 Generic functions may use traits as _bounds_ on their type parameters. This
1385 will have two effects:
1387 - Only types that have the trait may instantiate the parameter.
1388 - Within the generic function, the methods of the trait can be
1389 called on values that have the parameter's type.
1394 # type Surface = i32;
1395 # trait Shape { fn draw(&self, Surface); }
1396 fn draw_twice<T: Shape>(surface: Surface, sh: T) {
1402 Traits also define a [trait object](#trait-objects) with the same
1403 name as the trait. Values of this type are created by coercing from a
1404 pointer of some specific type to a pointer of trait type. For example,
1405 `&T` could be coerced to `&Shape` if `T: Shape` holds (and similarly
1406 for `Box<T>`). This coercion can either be implicit or
1407 [explicit](#type-cast-expressions). Here is an example of an explicit
1412 impl Shape for i32 { }
1413 let mycircle = 0i32;
1414 let myshape: Box<Shape> = Box::new(mycircle) as Box<Shape>;
1417 The resulting value is a box containing the value that was cast, along with
1418 information that identifies the methods of the implementation that was used.
1419 Values with a trait type can have [methods called](#method-call-expressions) on
1420 them, for any method in the trait, and can be used to instantiate type
1421 parameters that are bounded by the trait.
1423 Trait methods may be static, which means that they lack a `self` argument.
1424 This means that they can only be called with function call syntax (`f(x)`) and
1425 not method call syntax (`obj.f()`). The way to refer to the name of a static
1426 method is to qualify it with the trait name, treating the trait name like a
1427 module. For example:
1431 fn from_i32(n: i32) -> Self;
1434 fn from_i32(n: i32) -> f64 { n as f64 }
1436 let x: f64 = Num::from_i32(42);
1439 Traits may inherit from other traits. Consider the following example:
1442 trait Shape { fn area(&self) -> f64; }
1443 trait Circle : Shape { fn radius(&self) -> f64; }
1446 The syntax `Circle : Shape` means that types that implement `Circle` must also
1447 have an implementation for `Shape`. Multiple supertraits are separated by `+`,
1448 `trait Circle : Shape + PartialEq { }`. In an implementation of `Circle` for a
1449 given type `T`, methods can refer to `Shape` methods, since the typechecker
1450 checks that any type with an implementation of `Circle` also has an
1451 implementation of `Shape`:
1456 trait Shape { fn area(&self) -> f64; }
1457 trait Circle : Shape { fn radius(&self) -> f64; }
1458 impl Shape for Foo {
1459 fn area(&self) -> f64 {
1463 impl Circle for Foo {
1464 fn radius(&self) -> f64 {
1465 println!("calling area: {}", self.area());
1475 In type-parameterized functions, methods of the supertrait may be called on
1476 values of subtrait-bound type parameters. Referring to the previous example of
1477 `trait Circle : Shape`:
1480 # trait Shape { fn area(&self) -> f64; }
1481 # trait Circle : Shape { fn radius(&self) -> f64; }
1482 fn radius_times_area<T: Circle>(c: T) -> f64 {
1483 // `c` is both a Circle and a Shape
1484 c.radius() * c.area()
1488 Likewise, supertrait methods may also be called on trait objects.
1491 # trait Shape { fn area(&self) -> f64; }
1492 # trait Circle : Shape { fn radius(&self) -> f64; }
1493 # impl Shape for i32 { fn area(&self) -> f64 { 0.0 } }
1494 # impl Circle for i32 { fn radius(&self) -> f64 { 0.0 } }
1495 # let mycircle = 0i32;
1496 let mycircle = Box::new(mycircle) as Box<Circle>;
1497 let nonsense = mycircle.radius() * mycircle.area();
1502 An _implementation_ is an item that implements a [trait](#traits) for a
1505 Implementations are defined with the keyword `impl`.
1508 # #[derive(Copy, Clone)]
1509 # struct Point {x: f64, y: f64};
1510 # type Surface = i32;
1511 # struct BoundingBox {x: f64, y: f64, width: f64, height: f64};
1512 # trait Shape { fn draw(&self, Surface); fn bounding_box(&self) -> BoundingBox; }
1513 # fn do_draw_circle(s: Surface, c: Circle) { }
1519 impl Copy for Circle {}
1521 impl Clone for Circle {
1522 fn clone(&self) -> Circle { *self }
1525 impl Shape for Circle {
1526 fn draw(&self, s: Surface) { do_draw_circle(s, *self); }
1527 fn bounding_box(&self) -> BoundingBox {
1528 let r = self.radius;
1530 x: self.center.x - r,
1531 y: self.center.y - r,
1539 It is possible to define an implementation without referring to a trait. The
1540 methods in such an implementation can only be used as direct calls on the values
1541 of the type that the implementation targets. In such an implementation, the
1542 trait type and `for` after `impl` are omitted. Such implementations are limited
1543 to nominal types (enums, structs, trait objects), and the implementation must
1544 appear in the same crate as the `self` type:
1547 struct Point {x: i32, y: i32}
1551 println!("Point is at ({}, {})", self.x, self.y);
1555 let my_point = Point {x: 10, y:11};
1559 When a trait _is_ specified in an `impl`, all methods declared as part of the
1560 trait must be implemented, with matching types and type parameter counts.
1562 An implementation can take type parameters, which can be different from the
1563 type parameters taken by the trait it implements. Implementation parameters
1564 are written after the `impl` keyword.
1567 # trait Seq<T> { fn dummy(&self, _: T) { } }
1568 impl<T> Seq<T> for Vec<T> {
1571 impl Seq<bool> for u32 {
1572 /* Treat the integer as a sequence of bits */
1578 External blocks form the basis for Rust's foreign function interface.
1579 Declarations in an external block describe symbols in external, non-Rust
1582 Functions within external blocks are declared in the same way as other Rust
1583 functions, with the exception that they may not have a body and are instead
1584 terminated by a semicolon.
1586 Functions within external blocks may be called by Rust code, just like
1587 functions defined in Rust. The Rust compiler automatically translates between
1588 the Rust ABI and the foreign ABI.
1590 A number of [attributes](#attributes) control the behavior of external blocks.
1592 By default external blocks assume that the library they are calling uses the
1593 standard C "cdecl" ABI. Other ABIs may be specified using an `abi` string, as
1597 // Interface to the Windows API
1598 extern "stdcall" { }
1601 The `link` attribute allows the name of the library to be specified. When
1602 specified the compiler will attempt to link against the native library of the
1606 #[link(name = "crypto")]
1610 The type of a function declared in an extern block is `extern "abi" fn(A1, ...,
1611 An) -> R`, where `A1...An` are the declared types of its arguments and `R` is
1612 the declared return type.
1614 It is valid to add the `link` attribute on an empty extern block. You can use
1615 this to satisfy the linking requirements of extern blocks elsewhere in your code
1616 (including upstream crates) instead of adding the attribute to each extern block.
1618 ## Visibility and Privacy
1620 These two terms are often used interchangeably, and what they are attempting to
1621 convey is the answer to the question "Can this item be used at this location?"
1623 Rust's name resolution operates on a global hierarchy of namespaces. Each level
1624 in the hierarchy can be thought of as some item. The items are one of those
1625 mentioned above, but also include external crates. Declaring or defining a new
1626 module can be thought of as inserting a new tree into the hierarchy at the
1627 location of the definition.
1629 To control whether interfaces can be used across modules, Rust checks each use
1630 of an item to see whether it should be allowed or not. This is where privacy
1631 warnings are generated, or otherwise "you used a private item of another module
1632 and weren't allowed to."
1634 By default, everything in Rust is *private*, with one exception. Enum variants
1635 in a `pub` enum are also public by default. When an item is declared as `pub`,
1636 it can be thought of as being accessible to the outside world. For example:
1640 // Declare a private struct
1643 // Declare a public struct with a private field
1648 // Declare a public enum with two public variants
1650 PubliclyAccessibleState,
1651 PubliclyAccessibleState2,
1655 With the notion of an item being either public or private, Rust allows item
1656 accesses in two cases:
1658 1. If an item is public, then it can be used externally through any of its
1660 2. If an item is private, it may be accessed by the current module and its
1663 These two cases are surprisingly powerful for creating module hierarchies
1664 exposing public APIs while hiding internal implementation details. To help
1665 explain, here's a few use cases and what they would entail:
1667 * A library developer needs to expose functionality to crates which link
1668 against their library. As a consequence of the first case, this means that
1669 anything which is usable externally must be `pub` from the root down to the
1670 destination item. Any private item in the chain will disallow external
1673 * A crate needs a global available "helper module" to itself, but it doesn't
1674 want to expose the helper module as a public API. To accomplish this, the
1675 root of the crate's hierarchy would have a private module which then
1676 internally has a "public API". Because the entire crate is a descendant of
1677 the root, then the entire local crate can access this private module through
1680 * When writing unit tests for a module, it's often a common idiom to have an
1681 immediate child of the module to-be-tested named `mod test`. This module
1682 could access any items of the parent module through the second case, meaning
1683 that internal implementation details could also be seamlessly tested from the
1686 In the second case, it mentions that a private item "can be accessed" by the
1687 current module and its descendants, but the exact meaning of accessing an item
1688 depends on what the item is. Accessing a module, for example, would mean
1689 looking inside of it (to import more items). On the other hand, accessing a
1690 function would mean that it is invoked. Additionally, path expressions and
1691 import statements are considered to access an item in the sense that the
1692 import/expression is only valid if the destination is in the current visibility
1695 Here's an example of a program which exemplifies the three cases outlined
1699 // This module is private, meaning that no external crate can access this
1700 // module. Because it is private at the root of this current crate, however, any
1701 // module in the crate may access any publicly visible item in this module.
1702 mod crate_helper_module {
1704 // This function can be used by anything in the current crate
1705 pub fn crate_helper() {}
1707 // This function *cannot* be used by anything else in the crate. It is not
1708 // publicly visible outside of the `crate_helper_module`, so only this
1709 // current module and its descendants may access it.
1710 fn implementation_detail() {}
1713 // This function is "public to the root" meaning that it's available to external
1714 // crates linking against this one.
1715 pub fn public_api() {}
1717 // Similarly to 'public_api', this module is public so external crates may look
1720 use crate_helper_module;
1722 pub fn my_method() {
1723 // Any item in the local crate may invoke the helper module's public
1724 // interface through a combination of the two rules above.
1725 crate_helper_module::crate_helper();
1728 // This function is hidden to any module which is not a descendant of
1730 fn my_implementation() {}
1736 fn test_my_implementation() {
1737 // Because this module is a descendant of `submodule`, it's allowed
1738 // to access private items inside of `submodule` without a privacy
1740 super::my_implementation();
1748 For a rust program to pass the privacy checking pass, all paths must be valid
1749 accesses given the two rules above. This includes all use statements,
1750 expressions, types, etc.
1752 ### Re-exporting and Visibility
1754 Rust allows publicly re-exporting items through a `pub use` directive. Because
1755 this is a public directive, this allows the item to be used in the current
1756 module through the rules above. It essentially allows public access into the
1757 re-exported item. For example, this program is valid:
1760 pub use self::implementation::api;
1762 mod implementation {
1771 This means that any external crate referencing `implementation::api::f` would
1772 receive a privacy violation, while the path `api::f` would be allowed.
1774 When re-exporting a private item, it can be thought of as allowing the "privacy
1775 chain" being short-circuited through the reexport instead of passing through
1776 the namespace hierarchy as it normally would.
1780 Any item declaration may have an _attribute_ applied to it. Attributes in Rust
1781 are modeled on Attributes in ECMA-335, with the syntax coming from ECMA-334
1782 (C#). An attribute is a general, free-form metadatum that is interpreted
1783 according to name, convention, and language and compiler version. Attributes
1784 may appear as any of:
1786 * A single identifier, the attribute name
1787 * An identifier followed by the equals sign '=' and a literal, providing a
1789 * An identifier followed by a parenthesized list of sub-attribute arguments
1791 Attributes with a bang ("!") after the hash ("#") apply to the item that the
1792 attribute is declared within. Attributes that do not have a bang after the hash
1793 apply to the item that follows the attribute.
1795 An example of attributes:
1798 // General metadata applied to the enclosing module or crate.
1799 #![crate_type = "lib"]
1801 // A function marked as a unit test
1807 // A conditionally-compiled module
1808 #[cfg(target_os="linux")]
1813 // A lint attribute used to suppress a warning/error
1814 #[allow(non_camel_case_types)]
1818 > **Note:** At some point in the future, the compiler will distinguish between
1819 > language-reserved and user-available attributes. Until then, there is
1820 > effectively no difference between an attribute handled by a loadable syntax
1821 > extension and the compiler.
1823 ### Crate-only attributes
1825 - `crate_name` - specify the crate's crate name.
1826 - `crate_type` - see [linkage](#linkage).
1827 - `feature` - see [compiler features](#compiler-features).
1828 - `no_builtins` - disable optimizing certain code patterns to invocations of
1829 library functions that are assumed to exist
1830 - `no_main` - disable emitting the `main` symbol. Useful when some other
1831 object being linked to defines `main`.
1832 - `no_start` - disable linking to the `native` crate, which specifies the
1833 "start" language item.
1834 - `no_std` - disable linking to the `std` crate.
1835 - `plugin` - load a list of named crates as compiler plugins, e.g.
1836 `#![plugin(foo, bar)]`. Optional arguments for each plugin,
1837 i.e. `#![plugin(foo(... args ...))]`, are provided to the plugin's
1838 registrar function. The `plugin` feature gate is required to use
1840 - `recursion_limit` - Sets the maximum depth for potentially
1841 infinitely-recursive compile-time operations like
1842 auto-dereference or macro expansion. The default is
1843 `#![recursion_limit="64"]`.
1845 ### Module-only attributes
1847 - `no_implicit_prelude` - disable injecting `use std::prelude::*` in this
1849 - `path` - specifies the file to load the module from. `#[path="foo.rs"] mod
1850 bar;` is equivalent to `mod bar { /* contents of foo.rs */ }`. The path is
1851 taken relative to the directory that the current module is in.
1853 ### Function-only attributes
1855 - `main` - indicates that this function should be passed to the entry point,
1856 rather than the function in the crate root named `main`.
1857 - `plugin_registrar` - mark this function as the registration point for
1858 [compiler plugins][plugin], such as loadable syntax extensions.
1859 - `start` - indicates that this function should be used as the entry point,
1860 overriding the "start" language item. See the "start" [language
1861 item](#language-items) for more details.
1862 - `test` - indicates that this function is a test function, to only be compiled
1863 in case of `--test`.
1864 - `should_panic` - indicates that this test function should panic, inverting the success condition.
1865 - `cold` - The function is unlikely to be executed, so optimize it (and calls
1868 ### Static-only attributes
1870 - `thread_local` - on a `static mut`, this signals that the value of this
1871 static may change depending on the current thread. The exact consequences of
1872 this are implementation-defined.
1876 On an `extern` block, the following attributes are interpreted:
1878 - `link_args` - specify arguments to the linker, rather than just the library
1879 name and type. This is feature gated and the exact behavior is
1880 implementation-defined (due to variety of linker invocation syntax).
1881 - `link` - indicate that a native library should be linked to for the
1882 declarations in this block to be linked correctly. `link` supports an optional
1883 `kind` key with three possible values: `dylib`, `static`, and `framework`. See
1884 [external blocks](#external-blocks) for more about external blocks. Two
1885 examples: `#[link(name = "readline")]` and
1886 `#[link(name = "CoreFoundation", kind = "framework")]`.
1887 - `linked_from` - indicates what native library this block of FFI items is
1888 coming from. This attribute is of the form `#[linked_from = "foo"]` where
1889 `foo` is the name of a library in either `#[link]` or a `-l` flag. This
1890 attribute is currently required to export symbols from a Rust dynamic library
1891 on Windows, and it is feature gated behind the `linked_from` feature.
1893 On declarations inside an `extern` block, the following attributes are
1896 - `link_name` - the name of the symbol that this function or static should be
1898 - `linkage` - on a static, this specifies the [linkage
1899 type](http://llvm.org/docs/LangRef.html#linkage-types).
1903 - `repr` - on C-like enums, this sets the underlying type used for
1904 representation. Takes one argument, which is the primitive
1905 type this enum should be represented for, or `C`, which specifies that it
1906 should be the default `enum` size of the C ABI for that platform. Note that
1907 enum representation in C is undefined, and this may be incorrect when the C
1908 code is compiled with certain flags.
1912 - `repr` - specifies the representation to use for this struct. Takes a list
1913 of options. The currently accepted ones are `C` and `packed`, which may be
1914 combined. `C` will use a C ABI compatible struct layout, and `packed` will
1915 remove any padding between fields (note that this is very fragile and may
1916 break platforms which require aligned access).
1918 ### Macro-related attributes
1920 - `macro_use` on a `mod` — macros defined in this module will be visible in the
1921 module's parent, after this module has been included.
1923 - `macro_use` on an `extern crate` — load macros from this crate. An optional
1924 list of names `#[macro_use(foo, bar)]` restricts the import to just those
1925 macros named. The `extern crate` must appear at the crate root, not inside
1926 `mod`, which ensures proper function of the [`$crate` macro
1927 variable](book/macros.html#the-variable-crate).
1929 - `macro_reexport` on an `extern crate` — re-export the named macros.
1931 - `macro_export` - export a macro for cross-crate usage.
1933 - `no_link` on an `extern crate` — even if we load this crate for macros, don't
1934 link it into the output.
1936 See the [macros section of the
1937 book](book/macros.html#scoping-and-macro-importexport) for more information on
1941 ### Miscellaneous attributes
1943 - `export_name` - on statics and functions, this determines the name of the
1945 - `link_section` - on statics and functions, this specifies the section of the
1946 object file that this item's contents will be placed into.
1947 - `no_mangle` - on any item, do not apply the standard name mangling. Set the
1948 symbol for this item to its identifier.
1949 - `simd` - on certain tuple structs, derive the arithmetic operators, which
1950 lower to the target's SIMD instructions, if any; the `simd` feature gate
1951 is necessary to use this attribute.
1952 - `unsafe_destructor_blind_to_params` - on `Drop::drop` method, asserts that the
1953 destructor code (and all potential specializations of that code) will
1954 never attempt to read from nor write to any references with lifetimes
1955 that come in via generic parameters. This is a constraint we cannot
1956 currently express via the type system, and therefore we rely on the
1957 programmer to assert that it holds. Adding this to a Drop impl causes
1958 the associated destructor to be considered "uninteresting" by the
1959 Drop-Check rule, and thus it can help sidestep data ordering
1960 constraints that would otherwise be introduced by the Drop-Check
1961 rule. Such sidestepping of the constraints, if done incorrectly, can
1962 lead to undefined behavior (in the form of reading or writing to data
1963 outside of its dynamic extent), and thus this attribute has the word
1964 "unsafe" in its name. To use this, the
1965 `unsafe_destructor_blind_to_params` feature gate must be enabled.
1966 - `unsafe_no_drop_flag` - on structs, remove the flag that prevents
1967 destructors from being run twice. Destructors might be run multiple times on
1968 the same object with this attribute. To use this, the `unsafe_no_drop_flag` feature
1969 gate must be enabled.
1970 - `doc` - Doc comments such as `/// foo` are equivalent to `#[doc = "foo"]`.
1971 - `rustc_on_unimplemented` - Write a custom note to be shown along with the error
1972 when the trait is found to be unimplemented on a type.
1973 You may use format arguments like `{T}`, `{A}` to correspond to the
1974 types at the point of use corresponding to the type parameters of the
1975 trait of the same name. `{Self}` will be replaced with the type that is supposed
1976 to implement the trait but doesn't. To use this, the `on_unimplemented` feature gate
1979 ### Conditional compilation
1981 Sometimes one wants to have different compiler outputs from the same code,
1982 depending on build target, such as targeted operating system, or to enable
1985 There are two kinds of configuration options, one that is either defined or not
1986 (`#[cfg(foo)]`), and the other that contains a string that can be checked
1987 against (`#[cfg(bar = "baz")]`). Currently, only compiler-defined configuration
1988 options can have the latter form.
1991 // The function is only included in the build when compiling for OSX
1992 #[cfg(target_os = "macos")]
1997 // This function is only included when either foo or bar is defined
1998 #[cfg(any(foo, bar))]
1999 fn needs_foo_or_bar() {
2003 // This function is only included when compiling for a unixish OS with a 32-bit
2005 #[cfg(all(unix, target_pointer_width = "32"))]
2006 fn on_32bit_unix() {
2010 // This function is only included when foo is not defined
2012 fn needs_not_foo() {
2017 This illustrates some conditional compilation can be achieved using the
2018 `#[cfg(...)]` attribute. `any`, `all` and `not` can be used to assemble
2019 arbitrarily complex configurations through nesting.
2021 The following configurations must be defined by the implementation:
2023 * `debug_assertions` - Enabled by default when compiling without optimizations.
2024 This can be used to enable extra debugging code in development but not in
2025 production. For example, it controls the behavior of the standard library's
2026 `debug_assert!` macro.
2027 * `target_arch = "..."` - Target CPU architecture, such as `"x86"`, `"x86_64"`
2028 `"mips"`, `"powerpc"`, `"arm"`, or `"aarch64"`.
2029 * `target_endian = "..."` - Endianness of the target CPU, either `"little"` or
2031 * `target_env = ".."` - An option provided by the compiler by default
2032 describing the runtime environment of the target platform. Some examples of
2033 this are `musl` for builds targeting the MUSL libc implementation, `msvc` for
2034 Windows builds targeting MSVC, and `gnu` frequently the rest of the time. This
2035 option may also be blank on some platforms.
2036 * `target_family = "..."` - Operating system family of the target, e. g.
2037 `"unix"` or `"windows"`. The value of this configuration option is defined
2038 as a configuration itself, like `unix` or `windows`.
2039 * `target_os = "..."` - Operating system of the target, examples include
2040 `"windows"`, `"macos"`, `"ios"`, `"linux"`, `"android"`, `"freebsd"`, `"dragonfly"`,
2041 `"bitrig"` , `"openbsd"` or `"netbsd"`.
2042 * `target_pointer_width = "..."` - Target pointer width in bits. This is set
2043 to `"32"` for targets with 32-bit pointers, and likewise set to `"64"` for
2045 * `target_vendor = "..."` - Vendor of the target, for example `apple`, `pc`, or
2047 * `test` - Enabled when compiling the test harness (using the `--test` flag).
2048 * `unix` - See `target_family`.
2049 * `windows` - See `target_family`.
2051 You can also set another attribute based on a `cfg` variable with `cfg_attr`:
2057 Will be the same as `#[b]` if `a` is set by `cfg`, and nothing otherwise.
2059 ### Lint check attributes
2061 A lint check names a potentially undesirable coding pattern, such as
2062 unreachable code or omitted documentation, for the static entity to which the
2065 For any lint check `C`:
2067 * `allow(C)` overrides the check for `C` so that violations will go
2069 * `deny(C)` signals an error after encountering a violation of `C`,
2070 * `forbid(C)` is the same as `deny(C)`, but also forbids changing the lint
2072 * `warn(C)` warns about violations of `C` but continues compilation.
2074 The lint checks supported by the compiler can be found via `rustc -W help`,
2075 along with their default settings. [Compiler
2076 plugins](book/compiler-plugins.html#lint-plugins) can provide additional lint checks.
2080 // Missing documentation is ignored here
2081 #[allow(missing_docs)]
2082 pub fn undocumented_one() -> i32 { 1 }
2084 // Missing documentation signals a warning here
2085 #[warn(missing_docs)]
2086 pub fn undocumented_too() -> i32 { 2 }
2088 // Missing documentation signals an error here
2089 #[deny(missing_docs)]
2090 pub fn undocumented_end() -> i32 { 3 }
2094 This example shows how one can use `allow` and `warn` to toggle a particular
2098 #[warn(missing_docs)]
2100 #[allow(missing_docs)]
2102 // Missing documentation is ignored here
2103 pub fn undocumented_one() -> i32 { 1 }
2105 // Missing documentation signals a warning here,
2106 // despite the allow above.
2107 #[warn(missing_docs)]
2108 pub fn undocumented_two() -> i32 { 2 }
2111 // Missing documentation signals a warning here
2112 pub fn undocumented_too() -> i32 { 3 }
2116 This example shows how one can use `forbid` to disallow uses of `allow` for
2120 #[forbid(missing_docs)]
2122 // Attempting to toggle warning signals an error here
2123 #[allow(missing_docs)]
2125 pub fn undocumented_too() -> i32 { 2 }
2131 Some primitive Rust operations are defined in Rust code, rather than being
2132 implemented directly in C or assembly language. The definitions of these
2133 operations have to be easy for the compiler to find. The `lang` attribute
2134 makes it possible to declare these operations. For example, the `str` module
2135 in the Rust standard library defines the string equality function:
2139 pub fn eq_slice(a: &str, b: &str) -> bool {
2144 The name `str_eq` has a special meaning to the Rust compiler, and the presence
2145 of this definition means that it will use this definition when generating calls
2146 to the string equality function.
2148 The set of language items is currently considered unstable. A complete
2149 list of the built-in language items will be added in the future.
2151 ### Inline attributes
2153 The inline attribute suggests that the compiler should place a copy of
2154 the function or static in the caller, rather than generating code to
2155 call the function or access the static where it is defined.
2157 The compiler automatically inlines functions based on internal heuristics.
2158 Incorrectly inlining functions can actually make the program slower, so it
2159 should be used with care.
2161 `#[inline]` and `#[inline(always)]` always cause the function to be serialized
2162 into the crate metadata to allow cross-crate inlining.
2164 There are three different types of inline attributes:
2166 * `#[inline]` hints the compiler to perform an inline expansion.
2167 * `#[inline(always)]` asks the compiler to always perform an inline expansion.
2168 * `#[inline(never)]` asks the compiler to never perform an inline expansion.
2172 The `derive` attribute allows certain traits to be automatically implemented
2173 for data structures. For example, the following will create an `impl` for the
2174 `PartialEq` and `Clone` traits for `Foo`, the type parameter `T` will be given
2175 the `PartialEq` or `Clone` constraints for the appropriate `impl`:
2178 #[derive(PartialEq, Clone)]
2185 The generated `impl` for `PartialEq` is equivalent to
2188 # struct Foo<T> { a: i32, b: T }
2189 impl<T: PartialEq> PartialEq for Foo<T> {
2190 fn eq(&self, other: &Foo<T>) -> bool {
2191 self.a == other.a && self.b == other.b
2194 fn ne(&self, other: &Foo<T>) -> bool {
2195 self.a != other.a || self.b != other.b
2200 ### Compiler Features
2202 Certain aspects of Rust may be implemented in the compiler, but they're not
2203 necessarily ready for every-day use. These features are often of "prototype
2204 quality" or "almost production ready", but may not be stable enough to be
2205 considered a full-fledged language feature.
2207 For this reason, Rust recognizes a special crate-level attribute of the form:
2210 #![feature(feature1, feature2, feature3)]
2213 This directive informs the compiler that the feature list: `feature1`,
2214 `feature2`, and `feature3` should all be enabled. This is only recognized at a
2215 crate-level, not at a module-level. Without this directive, all features are
2216 considered off, and using the features will result in a compiler error.
2218 The currently implemented features of the reference compiler are:
2220 * `advanced_slice_patterns` - See the [match expressions](#match-expressions)
2221 section for discussion; the exact semantics of
2222 slice patterns are subject to change, so some types
2225 * `slice_patterns` - OK, actually, slice patterns are just scary and
2226 completely unstable.
2228 * `asm` - The `asm!` macro provides a means for inline assembly. This is often
2229 useful, but the exact syntax for this feature along with its
2230 semantics are likely to change, so this macro usage must be opted
2233 * `associated_consts` - Allows constants to be defined in `impl` and `trait`
2234 blocks, so that they can be associated with a type or
2235 trait in a similar manner to methods and associated
2238 * `box_patterns` - Allows `box` patterns, the exact semantics of which
2239 is subject to change.
2241 * `box_syntax` - Allows use of `box` expressions, the exact semantics of which
2242 is subject to change.
2244 * `cfg_target_vendor` - Allows conditional compilation using the `target_vendor`
2245 matcher which is subject to change.
2247 * `concat_idents` - Allows use of the `concat_idents` macro, which is in many
2248 ways insufficient for concatenating identifiers, and may be
2249 removed entirely for something more wholesome.
2251 * `custom_attribute` - Allows the usage of attributes unknown to the compiler
2252 so that new attributes can be added in a backwards compatible
2255 * `custom_derive` - Allows the use of `#[derive(Foo,Bar)]` as sugar for
2256 `#[derive_Foo] #[derive_Bar]`, which can be user-defined syntax
2259 * `intrinsics` - Allows use of the "rust-intrinsics" ABI. Compiler intrinsics
2260 are inherently unstable and no promise about them is made.
2262 * `lang_items` - Allows use of the `#[lang]` attribute. Like `intrinsics`,
2263 lang items are inherently unstable and no promise about them
2266 * `link_args` - This attribute is used to specify custom flags to the linker,
2267 but usage is strongly discouraged. The compiler's usage of the
2268 system linker is not guaranteed to continue in the future, and
2269 if the system linker is not used then specifying custom flags
2270 doesn't have much meaning.
2272 * `link_llvm_intrinsics` – Allows linking to LLVM intrinsics via
2273 `#[link_name="llvm.*"]`.
2275 * `linkage` - Allows use of the `linkage` attribute, which is not portable.
2277 * `log_syntax` - Allows use of the `log_syntax` macro attribute, which is a
2278 nasty hack that will certainly be removed.
2280 * `main` - Allows use of the `#[main]` attribute, which changes the entry point
2281 into a Rust program. This capability is subject to change.
2283 * `macro_reexport` - Allows macros to be re-exported from one crate after being imported
2284 from another. This feature was originally designed with the sole
2285 use case of the Rust standard library in mind, and is subject to
2288 * `non_ascii_idents` - The compiler supports the use of non-ascii identifiers,
2289 but the implementation is a little rough around the
2290 edges, so this can be seen as an experimental feature
2291 for now until the specification of identifiers is fully
2294 * `no_std` - Allows the `#![no_std]` crate attribute, which disables the implicit
2295 `extern crate std`. This typically requires use of the unstable APIs
2296 behind the libstd "facade", such as libcore and libcollections. It
2297 may also cause problems when using syntax extensions, including
2300 * `on_unimplemented` - Allows the `#[rustc_on_unimplemented]` attribute, which allows
2301 trait definitions to add specialized notes to error messages
2302 when an implementation was expected but not found.
2304 * `optin_builtin_traits` - Allows the definition of default and negative trait
2305 implementations. Experimental.
2307 * `plugin` - Usage of [compiler plugins][plugin] for custom lints or syntax extensions.
2308 These depend on compiler internals and are subject to change.
2310 * `plugin_registrar` - Indicates that a crate provides [compiler plugins][plugin].
2312 * `quote` - Allows use of the `quote_*!` family of macros, which are
2313 implemented very poorly and will likely change significantly
2314 with a proper implementation.
2316 * `rustc_attrs` - Gates internal `#[rustc_*]` attributes which may be
2317 for internal use only or have meaning added to them in the future.
2319 * `rustc_diagnostic_macros`- A mysterious feature, used in the implementation
2320 of rustc, not meant for mortals.
2322 * `simd` - Allows use of the `#[simd]` attribute, which is overly simple and
2323 not the SIMD interface we want to expose in the long term.
2325 * `simd_ffi` - Allows use of SIMD vectors in signatures for foreign functions.
2326 The SIMD interface is subject to change.
2328 * `start` - Allows use of the `#[start]` attribute, which changes the entry point
2329 into a Rust program. This capability, especially the signature for the
2330 annotated function, is subject to change.
2332 * `thread_local` - The usage of the `#[thread_local]` attribute is experimental
2333 and should be seen as unstable. This attribute is used to
2334 declare a `static` as being unique per-thread leveraging
2335 LLVM's implementation which works in concert with the kernel
2336 loader and dynamic linker. This is not necessarily available
2337 on all platforms, and usage of it is discouraged.
2339 * `trace_macros` - Allows use of the `trace_macros` macro, which is a nasty
2340 hack that will certainly be removed.
2342 * `unboxed_closures` - Rust's new closure design, which is currently a work in
2343 progress feature with many known bugs.
2345 * `unsafe_no_drop_flag` - Allows use of the `#[unsafe_no_drop_flag]` attribute,
2346 which removes hidden flag added to a type that
2347 implements the `Drop` trait. The design for the
2348 `Drop` flag is subject to change, and this feature
2349 may be removed in the future.
2351 * `unmarked_api` - Allows use of items within a `#![staged_api]` crate
2352 which have not been marked with a stability marker.
2353 Such items should not be allowed by the compiler to exist,
2354 so if you need this there probably is a compiler bug.
2356 * `visible_private_types` - Allows public APIs to expose otherwise private
2357 types, e.g. as the return type of a public function.
2358 This capability may be removed in the future.
2360 * `allow_internal_unstable` - Allows `macro_rules!` macros to be tagged with the
2361 `#[allow_internal_unstable]` attribute, designed
2362 to allow `std` macros to call
2363 `#[unstable]`/feature-gated functionality
2364 internally without imposing on callers
2365 (i.e. making them behave like function calls in
2366 terms of encapsulation).
2367 * - `default_type_parameter_fallback` - Allows type parameter defaults to
2368 influence type inference.
2369 * - `braced_empty_structs` - Allows use of empty structs and enum variants with braces.
2371 If a feature is promoted to a language feature, then all existing programs will
2372 start to receive compilation warnings about `#![feature]` directives which enabled
2373 the new feature (because the directive is no longer necessary). However, if a
2374 feature is decided to be removed from the language, errors will be issued (if
2375 there isn't a parser error first). The directive in this case is no longer
2376 necessary, and it's likely that existing code will break if the feature isn't
2379 If an unknown feature is found in a directive, it results in a compiler error.
2380 An unknown feature is one which has never been recognized by the compiler.
2382 # Statements and expressions
2384 Rust is _primarily_ an expression language. This means that most forms of
2385 value-producing or effect-causing evaluation are directed by the uniform syntax
2386 category of _expressions_. Each kind of expression can typically _nest_ within
2387 each other kind of expression, and rules for evaluation of expressions involve
2388 specifying both the value produced by the expression and the order in which its
2389 sub-expressions are themselves evaluated.
2391 In contrast, statements in Rust serve _mostly_ to contain and explicitly
2392 sequence expression evaluation.
2396 A _statement_ is a component of a block, which is in turn a component of an
2397 outer [expression](#expressions) or [function](#functions).
2399 Rust has two kinds of statement: [declaration
2400 statements](#declaration-statements) and [expression
2401 statements](#expression-statements).
2403 ### Declaration statements
2405 A _declaration statement_ is one that introduces one or more *names* into the
2406 enclosing statement block. The declared names may denote new variables or new
2409 #### Item declarations
2411 An _item declaration statement_ has a syntactic form identical to an
2412 [item](#items) declaration within a module. Declaring an item — a
2413 function, enumeration, struct, type, static, trait, implementation or module
2414 — locally within a statement block is simply a way of restricting its
2415 scope to a narrow region containing all of its uses; it is otherwise identical
2416 in meaning to declaring the item outside the statement block.
2418 > **Note**: there is no implicit capture of the function's dynamic environment when
2419 > declaring a function-local item.
2421 #### `let` statements
2423 A _`let` statement_ introduces a new set of variables, given by a pattern. The
2424 pattern may be followed by a type annotation, and/or an initializer expression.
2425 When no type annotation is given, the compiler will infer the type, or signal
2426 an error if insufficient type information is available for definite inference.
2427 Any variables introduced by a variable declaration are visible from the point of
2428 declaration until the end of the enclosing block scope.
2430 ### Expression statements
2432 An _expression statement_ is one that evaluates an [expression](#expressions)
2433 and ignores its result. The type of an expression statement `e;` is always
2434 `()`, regardless of the type of `e`. As a rule, an expression statement's
2435 purpose is to trigger the effects of evaluating its expression.
2439 An expression may have two roles: it always produces a *value*, and it may have
2440 *effects* (otherwise known as "side effects"). An expression *evaluates to* a
2441 value, and has effects during *evaluation*. Many expressions contain
2442 sub-expressions (operands). The meaning of each kind of expression dictates
2445 * Whether or not to evaluate the sub-expressions when evaluating the expression
2446 * The order in which to evaluate the sub-expressions
2447 * How to combine the sub-expressions' values to obtain the value of the expression
2449 In this way, the structure of expressions dictates the structure of execution.
2450 Blocks are just another kind of expression, so blocks, statements, expressions,
2451 and blocks again can recursively nest inside each other to an arbitrary depth.
2453 #### Lvalues, rvalues and temporaries
2455 Expressions are divided into two main categories: _lvalues_ and _rvalues_.
2456 Likewise within each expression, sub-expressions may occur in _lvalue context_
2457 or _rvalue context_. The evaluation of an expression depends both on its own
2458 category and the context it occurs within.
2460 An lvalue is an expression that represents a memory location. These expressions
2461 are [paths](#path-expressions) (which refer to local variables, function and
2462 method arguments, or static variables), dereferences (`*expr`), [indexing
2463 expressions](#index-expressions) (`expr[expr]`), and [field
2464 references](#field-expressions) (`expr.f`). All other expressions are rvalues.
2466 The left operand of an [assignment](#assignment-expressions) or
2467 [compound-assignment](#compound-assignment-expressions) expression is
2468 an lvalue context, as is the single operand of a unary
2469 [borrow](#unary-operator-expressions). The discriminant or subject of
2470 a [match expression](#match-expressions) may be an lvalue context, if
2471 ref bindings are made, but is otherwise an rvalue context. All other
2472 expression contexts are rvalue contexts.
2474 When an lvalue is evaluated in an _lvalue context_, it denotes a memory
2475 location; when evaluated in an _rvalue context_, it denotes the value held _in_
2476 that memory location.
2478 ##### Temporary lifetimes
2480 When an rvalue is used in an lvalue context, a temporary un-named
2481 lvalue is created and used instead. The lifetime of temporary values
2482 is typically the innermost enclosing statement; the tail expression of
2483 a block is considered part of the statement that encloses the block.
2485 When a temporary rvalue is being created that is assigned into a `let`
2486 declaration, however, the temporary is created with the lifetime of
2487 the enclosing block instead, as using the enclosing statement (the
2488 `let` declaration) would be a guaranteed error (since a pointer to the
2489 temporary would be stored into a variable, but the temporary would be
2490 freed before the variable could be used). The compiler uses simple
2491 syntactic rules to decide which values are being assigned into a `let`
2492 binding, and therefore deserve a longer temporary lifetime.
2494 Here are some examples:
2496 - `let x = foo(&temp())`. The expression `temp()` is an rvalue. As it
2497 is being borrowed, a temporary is created which will be freed after
2498 the innermost enclosing statement (the `let` declaration, in this case).
2499 - `let x = temp().foo()`. This is the same as the previous example,
2500 except that the value of `temp()` is being borrowed via autoref on a
2501 method-call. Here we are assuming that `foo()` is an `&self` method
2502 defined in some trait, say `Foo`. In other words, the expression
2503 `temp().foo()` is equivalent to `Foo::foo(&temp())`.
2504 - `let x = &temp()`. Here, the same temporary is being assigned into
2505 `x`, rather than being passed as a parameter, and hence the
2506 temporary's lifetime is considered to be the enclosing block.
2507 - `let x = SomeStruct { foo: &temp() }`. As in the previous case, the
2508 temporary is assigned into a struct which is then assigned into a
2509 binding, and hence it is given the lifetime of the enclosing block.
2510 - `let x = [ &temp() ]`. As in the previous case, the
2511 temporary is assigned into an array which is then assigned into a
2512 binding, and hence it is given the lifetime of the enclosing block.
2513 - `let ref x = temp()`. In this case, the temporary is created using a ref binding,
2514 but the result is the same: the lifetime is extended to the enclosing block.
2516 #### Moved and copied types
2518 When a [local variable](#variables) is used as an
2519 [rvalue](#lvalues-rvalues-and-temporaries), the variable will be copied
2520 if its type implements `Copy`. All others are moved.
2522 ### Literal expressions
2524 A _literal expression_ consists of one of the [literal](#literals) forms
2525 described earlier. It directly describes a number, character, string, boolean
2526 value, or the unit value.
2530 "hello"; // string type
2531 '5'; // character type
2535 ### Path expressions
2537 A [path](#paths) used as an expression context denotes either a local variable
2538 or an item. Path expressions are [lvalues](#lvalues-rvalues-and-temporaries).
2540 ### Tuple expressions
2542 Tuples are written by enclosing zero or more comma-separated expressions in
2543 parentheses. They are used to create [tuple-typed](#tuple-types) values.
2547 ("a", 4usize, true);
2550 You can disambiguate a single-element tuple from a value in parentheses with a
2554 (0,); // single-element tuple
2555 (0); // zero in parentheses
2558 ### Struct expressions
2560 There are several forms of struct expressions. A _struct expression_
2561 consists of the [path](#paths) of a [struct item](#structs), followed by
2562 a brace-enclosed list of one or more comma-separated name-value pairs,
2563 providing the field values of a new instance of the struct. A field name
2564 can be any identifier, and is separated from its value expression by a colon.
2565 The location denoted by a struct field is mutable if and only if the
2566 enclosing struct is mutable.
2568 A _tuple struct expression_ consists of the [path](#paths) of a [struct
2569 item](#structs), followed by a parenthesized list of one or more
2570 comma-separated expressions (in other words, the path of a struct item
2571 followed by a tuple expression). The struct item must be a tuple struct
2574 A _unit-like struct expression_ consists only of the [path](#paths) of a
2575 [struct item](#structs).
2577 The following are examples of struct expressions:
2580 # struct Point { x: f64, y: f64 }
2581 # struct TuplePoint(f64, f64);
2582 # mod game { pub struct User<'a> { pub name: &'a str, pub age: u32, pub score: usize } }
2583 # struct Cookie; fn some_fn<T>(t: T) {}
2584 Point {x: 10.0, y: 20.0};
2585 TuplePoint(10.0, 20.0);
2586 let u = game::User {name: "Joe", age: 35, score: 100_000};
2587 some_fn::<Cookie>(Cookie);
2590 A struct expression forms a new value of the named struct type. Note
2591 that for a given *unit-like* struct type, this will always be the same
2594 A struct expression can terminate with the syntax `..` followed by an
2595 expression to denote a functional update. The expression following `..` (the
2596 base) must have the same struct type as the new struct type being formed.
2597 The entire expression denotes the result of constructing a new struct (with
2598 the same type as the base expression) with the given values for the fields that
2599 were explicitly specified and the values in the base expression for all other
2603 # struct Point3d { x: i32, y: i32, z: i32 }
2604 let base = Point3d {x: 1, y: 2, z: 3};
2605 Point3d {y: 0, z: 10, .. base};
2608 ### Block expressions
2610 A _block expression_ is similar to a module in terms of the declarations that
2611 are possible. Each block conceptually introduces a new namespace scope. Use
2612 items can bring new names into scopes and declared items are in scope for only
2615 A block will execute each statement sequentially, and then execute the
2616 expression (if given). If the block ends in a statement, its value is `()`:
2619 let x: () = { println!("Hello."); };
2622 If it ends in an expression, its value and type are that of the expression:
2625 let x: i32 = { println!("Hello."); 5 };
2630 ### Method-call expressions
2632 A _method call_ consists of an expression followed by a single dot, an
2633 identifier, and a parenthesized expression-list. Method calls are resolved to
2634 methods on specific traits, either statically dispatching to a method if the
2635 exact `self`-type of the left-hand-side is known, or dynamically dispatching if
2636 the left-hand-side expression is an indirect [trait object](#trait-objects).
2638 ### Field expressions
2640 A _field expression_ consists of an expression followed by a single dot and an
2641 identifier, when not immediately followed by a parenthesized expression-list
2642 (the latter is a [method call expression](#method-call-expressions)). A field
2643 expression denotes a field of a [struct](#struct-types).
2648 (Struct {a: 10, b: 20}).a;
2651 A field access is an [lvalue](#lvalues-rvalues-and-temporaries) referring to
2652 the value of that field. When the type providing the field inherits mutability,
2653 it can be [assigned](#assignment-expressions) to.
2655 Also, if the type of the expression to the left of the dot is a
2656 pointer, it is automatically dereferenced as many times as necessary
2657 to make the field access possible. In cases of ambiguity, we prefer
2658 fewer autoderefs to more.
2660 ### Array expressions
2662 An [array](#array-and-slice-types) _expression_ is written by enclosing zero
2663 or more comma-separated expressions of uniform type in square brackets.
2665 In the `[expr ';' expr]` form, the expression after the `';'` must be a
2666 constant expression that can be evaluated at compile time, such as a
2667 [literal](#literals) or a [static item](#static-items).
2671 ["a", "b", "c", "d"];
2672 [0; 128]; // array with 128 zeros
2673 [0u8, 0u8, 0u8, 0u8];
2676 ### Index expressions
2678 [Array](#array-and-slice-types)-typed expressions can be indexed by
2679 writing a square-bracket-enclosed expression (the index) after them. When the
2680 array is mutable, the resulting [lvalue](#lvalues-rvalues-and-temporaries) can
2683 Indices are zero-based, and may be of any integral type. Vector access is
2684 bounds-checked at compile-time for constant arrays being accessed with a constant index value.
2685 Otherwise a check will be performed at run-time that will put the thread in a _panicked state_ if it fails.
2690 let x = (["a", "b"])[10]; // compiler error: const index-expr is out of bounds
2693 let y = (["a", "b"])[n]; // panics
2695 let arr = ["a", "b"];
2699 Also, if the type of the expression to the left of the brackets is a
2700 pointer, it is automatically dereferenced as many times as necessary
2701 to make the indexing possible. In cases of ambiguity, we prefer fewer
2704 ### Range expressions
2706 The `..` operator will construct an object of one of the `std::ops::Range` variants.
2709 1..2; // std::ops::Range
2710 3..; // std::ops::RangeFrom
2711 ..4; // std::ops::RangeTo
2712 ..; // std::ops::RangeFull
2715 The following expressions are equivalent.
2718 let x = std::ops::Range {start: 0, end: 10};
2724 ### Unary operator expressions
2726 Rust defines the following unary operators. They are all written as prefix operators,
2727 before the expression they apply to.
2730 : Negation. May only be applied to numeric types.
2732 : Dereference. When applied to a [pointer](#pointer-types) it denotes the
2733 pointed-to location. For pointers to mutable locations, the resulting
2734 [lvalue](#lvalues-rvalues-and-temporaries) can be assigned to.
2735 On non-pointer types, it calls the `deref` method of the `std::ops::Deref`
2736 trait, or the `deref_mut` method of the `std::ops::DerefMut` trait (if
2737 implemented by the type and required for an outer expression that will or
2738 could mutate the dereference), and produces the result of dereferencing the
2739 `&` or `&mut` borrowed pointer returned from the overload method.
2741 : Logical negation. On the boolean type, this flips between `true` and
2742 `false`. On integer types, this inverts the individual bits in the
2743 two's complement representation of the value.
2745 : Borrowing. When applied to an lvalue, these operators produce a
2746 reference (pointer) to the lvalue. The lvalue is also placed into
2747 a borrowed state for the duration of the reference. For a shared
2748 borrow (`&`), this implies that the lvalue may not be mutated, but
2749 it may be read or shared again. For a mutable borrow (`&mut`), the
2750 lvalue may not be accessed in any way until the borrow expires.
2751 If the `&` or `&mut` operators are applied to an rvalue, a
2752 temporary value is created; the lifetime of this temporary value
2753 is defined by [syntactic rules](#temporary-lifetimes).
2755 ### Binary operator expressions
2757 Binary operators expressions are given in terms of [operator
2758 precedence](#operator-precedence).
2760 #### Arithmetic operators
2762 Binary arithmetic expressions are syntactic sugar for calls to built-in traits,
2763 defined in the `std::ops` module of the `std` library. This means that
2764 arithmetic operators can be overridden for user-defined types. The default
2765 meaning of the operators on standard types is given here.
2768 : Addition and array/string concatenation.
2769 Calls the `add` method on the `std::ops::Add` trait.
2772 Calls the `sub` method on the `std::ops::Sub` trait.
2775 Calls the `mul` method on the `std::ops::Mul` trait.
2778 Calls the `div` method on the `std::ops::Div` trait.
2781 Calls the `rem` method on the `std::ops::Rem` trait.
2783 #### Bitwise operators
2785 Like the [arithmetic operators](#arithmetic-operators), bitwise operators are
2786 syntactic sugar for calls to methods of built-in traits. This means that
2787 bitwise operators can be overridden for user-defined types. The default
2788 meaning of the operators on standard types is given here. Bitwise `&`, `|` and
2789 `^` applied to boolean arguments are equivalent to logical `&&`, `||` and `!=`
2790 evaluated in non-lazy fashion.
2794 Calls the `bitand` method of the `std::ops::BitAnd` trait.
2796 : Bitwise inclusive OR.
2797 Calls the `bitor` method of the `std::ops::BitOr` trait.
2799 : Bitwise exclusive OR.
2800 Calls the `bitxor` method of the `std::ops::BitXor` trait.
2803 Calls the `shl` method of the `std::ops::Shl` trait.
2805 : Right shift (arithmetic).
2806 Calls the `shr` method of the `std::ops::Shr` trait.
2808 #### Lazy boolean operators
2810 The operators `||` and `&&` may be applied to operands of boolean type. The
2811 `||` operator denotes logical 'or', and the `&&` operator denotes logical
2812 'and'. They differ from `|` and `&` in that the right-hand operand is only
2813 evaluated when the left-hand operand does not already determine the result of
2814 the expression. That is, `||` only evaluates its right-hand operand when the
2815 left-hand operand evaluates to `false`, and `&&` only when it evaluates to
2818 #### Comparison operators
2820 Comparison operators are, like the [arithmetic
2821 operators](#arithmetic-operators), and [bitwise operators](#bitwise-operators),
2822 syntactic sugar for calls to built-in traits. This means that comparison
2823 operators can be overridden for user-defined types. The default meaning of the
2824 operators on standard types is given here.
2828 Calls the `eq` method on the `std::cmp::PartialEq` trait.
2831 Calls the `ne` method on the `std::cmp::PartialEq` trait.
2834 Calls the `lt` method on the `std::cmp::PartialOrd` trait.
2837 Calls the `gt` method on the `std::cmp::PartialOrd` trait.
2839 : Less than or equal.
2840 Calls the `le` method on the `std::cmp::PartialOrd` trait.
2842 : Greater than or equal.
2843 Calls the `ge` method on the `std::cmp::PartialOrd` trait.
2845 #### Type cast expressions
2847 A type cast expression is denoted with the binary operator `as`.
2849 Executing an `as` expression casts the value on the left-hand side to the type
2850 on the right-hand side.
2852 An example of an `as` expression:
2855 # fn sum(values: &[f64]) -> f64 { 0.0 }
2856 # fn len(values: &[f64]) -> i32 { 0 }
2858 fn average(values: &[f64]) -> f64 {
2859 let sum: f64 = sum(values);
2860 let size: f64 = len(values) as f64;
2865 Some of the conversions which can be done through the `as` operator
2866 can also be done implicitly at various points in the program, such as
2867 argument passing and assignment to a `let` binding with an explicit
2868 type. Implicit conversions are limited to "harmless" conversions that
2869 do not lose information and which have minimal or no risk of
2870 surprising side-effects on the dynamic execution semantics.
2872 #### Assignment expressions
2874 An _assignment expression_ consists of an
2875 [lvalue](#lvalues-rvalues-and-temporaries) expression followed by an equals
2876 sign (`=`) and an [rvalue](#lvalues-rvalues-and-temporaries) expression.
2878 Evaluating an assignment expression [either copies or
2879 moves](#moved-and-copied-types) its right-hand operand to its left-hand
2888 #### Compound assignment expressions
2890 The `+`, `-`, `*`, `/`, `%`, `&`, `|`, `^`, `<<`, and `>>` operators may be
2891 composed with the `=` operator. The expression `lval OP= val` is equivalent to
2892 `lval = lval OP val`. For example, `x = x + 1` may be written as `x += 1`.
2894 Any such expression always has the [`unit`](#tuple-types) type.
2896 #### Operator precedence
2898 The precedence of Rust binary operators is ordered as follows, going from
2901 ```{.text .precedence}
2915 Operators at the same precedence level are evaluated left-to-right. [Unary
2916 operators](#unary-operator-expressions) have the same precedence level and are
2917 stronger than any of the binary operators.
2919 ### Grouped expressions
2921 An expression enclosed in parentheses evaluates to the result of the enclosed
2922 expression. Parentheses can be used to explicitly specify evaluation order
2923 within an expression.
2925 An example of a parenthesized expression:
2928 let x: i32 = (2 + 3) * 4;
2932 ### Call expressions
2934 A _call expression_ invokes a function, providing zero or more input variables
2935 and an optional location to move the function's output into. If the function
2936 eventually returns, then the expression completes.
2938 Some examples of call expressions:
2941 # fn add(x: i32, y: i32) -> i32 { 0 }
2943 let x: i32 = add(1i32, 2i32);
2944 let pi: Result<f32, _> = "3.14".parse();
2947 ### Lambda expressions
2949 A _lambda expression_ (sometimes called an "anonymous function expression")
2950 defines a function and denotes it as a value, in a single expression. A lambda
2951 expression is a pipe-symbol-delimited (`|`) list of identifiers followed by an
2954 A lambda expression denotes a function that maps a list of parameters
2955 (`ident_list`) onto the expression that follows the `ident_list`. The
2956 identifiers in the `ident_list` are the parameters to the function. These
2957 parameters' types need not be specified, as the compiler infers them from
2960 Lambda expressions are most useful when passing functions as arguments to other
2961 functions, as an abbreviation for defining and capturing a separate function.
2963 Significantly, lambda expressions _capture their environment_, which regular
2964 [function definitions](#functions) do not. The exact type of capture depends
2965 on the [function type](#function-types) inferred for the lambda expression. In
2966 the simplest and least-expensive form (analogous to a ```|| { }``` expression),
2967 the lambda expression captures its environment by reference, effectively
2968 borrowing pointers to all outer variables mentioned inside the function.
2969 Alternately, the compiler may infer that a lambda expression should copy or
2970 move values (depending on their type) from the environment into the lambda
2971 expression's captured environment.
2973 In this example, we define a function `ten_times` that takes a higher-order
2974 function argument, and we then call it with a lambda expression as an argument:
2977 fn ten_times<F>(f: F) where F: Fn(i32) {
2978 for index in 0..10 {
2983 ten_times(|j| println!("hello, {}", j));
2988 A `loop` expression denotes an infinite loop.
2990 A `loop` expression may optionally have a _label_. The label is written as
2991 a lifetime preceding the loop expression, as in `'foo: loop{ }`. If a
2992 label is present, then labeled `break` and `continue` expressions nested
2993 within this loop may exit out of this loop or return control to its head.
2994 See [break expressions](#break-expressions) and [continue
2995 expressions](#continue-expressions).
2997 ### `break` expressions
2999 A `break` expression has an optional _label_. If the label is absent, then
3000 executing a `break` expression immediately terminates the innermost loop
3001 enclosing it. It is only permitted in the body of a loop. If the label is
3002 present, then `break 'foo` terminates the loop with label `'foo`, which need not
3003 be the innermost label enclosing the `break` expression, but must enclose it.
3005 ### `continue` expressions
3007 A `continue` expression has an optional _label_. If the label is absent, then
3008 executing a `continue` expression immediately terminates the current iteration
3009 of the innermost loop enclosing it, returning control to the loop *head*. In
3010 the case of a `while` loop, the head is the conditional expression controlling
3011 the loop. In the case of a `for` loop, the head is the call-expression
3012 controlling the loop. If the label is present, then `continue 'foo` returns
3013 control to the head of the loop with label `'foo`, which need not be the
3014 innermost label enclosing the `break` expression, but must enclose it.
3016 A `continue` expression is only permitted in the body of a loop.
3020 A `while` loop begins by evaluating the boolean loop conditional expression.
3021 If the loop conditional expression evaluates to `true`, the loop body block
3022 executes and control returns to the loop conditional expression. If the loop
3023 conditional expression evaluates to `false`, the `while` expression completes.
3036 Like `loop` expressions, `while` loops can be controlled with `break` or
3037 `continue`, and may optionally have a _label_. See [infinite
3038 loops](#infinite-loops), [break expressions](#break-expressions), and
3039 [continue expressions](#continue-expressions) for more information.
3041 ### `for` expressions
3043 A `for` expression is a syntactic construct for looping over elements provided
3044 by an implementation of `std::iter::IntoIterator`.
3046 An example of a `for` loop over the contents of an array:
3050 # fn bar(f: &Foo) { }
3055 let v: &[Foo] = &[a, b, c];
3062 An example of a for loop over a series of integers:
3065 # fn bar(b:usize) { }
3071 Like `loop` expressions, `for` loops can be controlled with `break` or
3072 `continue`, and may optionally have a _label_. See [infinite
3073 loops](#infinite-loops), [break expressions](#break-expressions), and
3074 [continue expressions](#continue-expressions) for more information.
3076 ### `if` expressions
3078 An `if` expression is a conditional branch in program control. The form of an
3079 `if` expression is a condition expression, followed by a consequent block, any
3080 number of `else if` conditions and blocks, and an optional trailing `else`
3081 block. The condition expressions must have type `bool`. If a condition
3082 expression evaluates to `true`, the consequent block is executed and any
3083 subsequent `else if` or `else` block is skipped. If a condition expression
3084 evaluates to `false`, the consequent block is skipped and any subsequent `else
3085 if` condition is evaluated. If all `if` and `else if` conditions evaluate to
3086 `false` then any `else` block is executed.
3088 ### `match` expressions
3090 A `match` expression branches on a *pattern*. The exact form of matching that
3091 occurs depends on the pattern. Patterns consist of some combination of
3092 literals, destructured arrays or enum constructors, structs and tuples,
3093 variable binding specifications, wildcards (`..`), and placeholders (`_`). A
3094 `match` expression has a *head expression*, which is the value to compare to
3095 the patterns. The type of the patterns must equal the type of the head
3098 In a pattern whose head expression has an `enum` type, a placeholder (`_`)
3099 stands for a *single* data field, whereas a wildcard `..` stands for *all* the
3100 fields of a particular variant.
3102 A `match` behaves differently depending on whether or not the head expression
3103 is an [lvalue or an rvalue](#lvalues-rvalues-and-temporaries). If the head
3104 expression is an rvalue, it is first evaluated into a temporary location, and
3105 the resulting value is sequentially compared to the patterns in the arms until
3106 a match is found. The first arm with a matching pattern is chosen as the branch
3107 target of the `match`, any variables bound by the pattern are assigned to local
3108 variables in the arm's block, and control enters the block.
3110 When the head expression is an lvalue, the match does not allocate a temporary
3111 location (however, a by-value binding may copy or move from the lvalue). When
3112 possible, it is preferable to match on lvalues, as the lifetime of these
3113 matches inherits the lifetime of the lvalue, rather than being restricted to
3114 the inside of the match.
3116 An example of a `match` expression:
3122 1 => println!("one"),
3123 2 => println!("two"),
3124 3 => println!("three"),
3125 4 => println!("four"),
3126 5 => println!("five"),
3127 _ => println!("something else"),
3131 Patterns that bind variables default to binding to a copy or move of the
3132 matched value (depending on the matched value's type). This can be changed to
3133 bind to a reference by using the `ref` keyword, or to a mutable reference using
3136 Subpatterns can also be bound to variables by the use of the syntax `variable @
3137 subpattern`. For example:
3143 e @ 1 ... 5 => println!("got a range element {}", e),
3144 _ => println!("anything"),
3148 Patterns can also dereference pointers by using the `&`, `&mut` and `box`
3149 symbols, as appropriate. For example, these two matches on `x: &i32` are
3154 let y = match *x { 0 => "zero", _ => "some" };
3155 let z = match x { &0 => "zero", _ => "some" };
3160 Multiple match patterns may be joined with the `|` operator. A range of values
3161 may be specified with `...`. For example:
3166 let message = match x {
3167 0 | 1 => "not many",
3173 Range patterns only work on scalar types (like integers and characters; not
3174 like arrays and structs, which have sub-components). A range pattern may not
3175 be a sub-range of another range pattern inside the same `match`.
3177 Finally, match patterns can accept *pattern guards* to further refine the
3178 criteria for matching a case. Pattern guards appear after the pattern and
3179 consist of a bool-typed expression following the `if` keyword. A pattern guard
3180 may refer to the variables bound within the pattern they follow.
3183 # let maybe_digit = Some(0);
3184 # fn process_digit(i: i32) { }
3185 # fn process_other(i: i32) { }
3187 let message = match maybe_digit {
3188 Some(x) if x < 10 => process_digit(x),
3189 Some(x) => process_other(x),
3194 ### `if let` expressions
3196 An `if let` expression is semantically identical to an `if` expression but in
3197 place of a condition expression it expects a `let` statement with a refutable
3198 pattern. If the value of the expression on the right hand side of the `let`
3199 statement matches the pattern, the corresponding block will execute, otherwise
3200 flow proceeds to the first `else` block that follows.
3203 let dish = ("Ham", "Eggs");
3205 // this body will be skipped because the pattern is refuted
3206 if let ("Bacon", b) = dish {
3207 println!("Bacon is served with {}", b);
3210 // this body will execute
3211 if let ("Ham", b) = dish {
3212 println!("Ham is served with {}", b);
3216 ### `while let` loops
3218 A `while let` loop is semantically identical to a `while` loop but in place of
3219 a condition expression it expects `let` statement with a refutable pattern. If
3220 the value of the expression on the right hand side of the `let` statement
3221 matches the pattern, the loop body block executes and control returns to the
3222 pattern matching statement. Otherwise, the while expression completes.
3224 ### `return` expressions
3226 Return expressions are denoted with the keyword `return`. Evaluating a `return`
3227 expression moves its argument into the designated output location for the
3228 current function call, destroys the current function activation frame, and
3229 transfers control to the caller frame.
3231 An example of a `return` expression:
3234 fn max(a: i32, b: i32) -> i32 {
3246 Every variable, item and value in a Rust program has a type. The _type_ of a
3247 *value* defines the interpretation of the memory holding it.
3249 Built-in types and type-constructors are tightly integrated into the language,
3250 in nontrivial ways that are not possible to emulate in user-defined types.
3251 User-defined types have limited capabilities.
3255 The primitive types are the following:
3257 * The boolean type `bool` with values `true` and `false`.
3258 * The machine types (integer and floating-point).
3259 * The machine-dependent integer types.
3263 The machine types are the following:
3265 * The unsigned word types `u8`, `u16`, `u32` and `u64`, with values drawn from
3266 the integer intervals [0, 2^8 - 1], [0, 2^16 - 1], [0, 2^32 - 1] and
3267 [0, 2^64 - 1] respectively.
3269 * The signed two's complement word types `i8`, `i16`, `i32` and `i64`, with
3270 values drawn from the integer intervals [-(2^(7)), 2^7 - 1],
3271 [-(2^(15)), 2^15 - 1], [-(2^(31)), 2^31 - 1], [-(2^(63)), 2^63 - 1]
3274 * The IEEE 754-2008 `binary32` and `binary64` floating-point types: `f32` and
3275 `f64`, respectively.
3277 #### Machine-dependent integer types
3279 The `usize` type is an unsigned integer type with the same number of bits as the
3280 platform's pointer type. It can represent every memory address in the process.
3282 The `isize` type is a signed integer type with the same number of bits as the
3283 platform's pointer type. The theoretical upper bound on object and array size
3284 is the maximum `isize` value. This ensures that `isize` can be used to calculate
3285 differences between pointers into an object or array and can address every byte
3286 within an object along with one byte past the end.
3290 The types `char` and `str` hold textual data.
3292 A value of type `char` is a [Unicode scalar value](
3293 http://www.unicode.org/glossary/#unicode_scalar_value) (i.e. a code point that
3294 is not a surrogate), represented as a 32-bit unsigned word in the 0x0000 to
3295 0xD7FF or 0xE000 to 0x10FFFF range. A `[char]` array is effectively an UCS-4 /
3298 A value of type `str` is a Unicode string, represented as an array of 8-bit
3299 unsigned bytes holding a sequence of UTF-8 code points. Since `str` is of
3300 unknown size, it is not a _first-class_ type, but can only be instantiated
3301 through a pointer type, such as `&str`.
3305 A tuple *type* is a heterogeneous product of other types, called the *elements*
3306 of the tuple. It has no nominal name and is instead structurally typed.
3308 Tuple types and values are denoted by listing the types or values of their
3309 elements, respectively, in a parenthesized, comma-separated list.
3311 Because tuple elements don't have a name, they can only be accessed by
3312 pattern-matching or by using `N` directly as a field to access the
3315 An example of a tuple type and its use:
3318 type Pair<'a> = (i32, &'a str);
3319 let p: Pair<'static> = (10, "ten");
3323 assert_eq!(b, "ten");
3324 assert_eq!(p.0, 10);
3325 assert_eq!(p.1, "ten");
3328 For historical reasons and convenience, the tuple type with no elements (`()`)
3329 is often called ‘unit’ or ‘the unit type’.
3331 ### Array, and Slice types
3333 Rust has two different types for a list of items:
3335 * `[T; N]`, an 'array'
3338 An array has a fixed size, and can be allocated on either the stack or the
3341 A slice is a 'view' into an array. It doesn't own the data it points
3347 // A stack-allocated array
3348 let array: [i32; 3] = [1, 2, 3];
3350 // A heap-allocated array
3351 let vector: Vec<i32> = vec![1, 2, 3];
3353 // A slice into an array
3354 let slice: &[i32] = &vector[..];
3357 As you can see, the `vec!` macro allows you to create a `Vec<T>` easily. The
3358 `vec!` macro is also part of the standard library, rather than the language.
3360 All in-bounds elements of arrays and slices are always initialized, and access
3361 to an array or slice is always bounds-checked.
3365 A `struct` *type* is a heterogeneous product of other types, called the
3366 *fields* of the type.[^structtype]
3368 [^structtype]: `struct` types are analogous to `struct` types in C,
3369 the *record* types of the ML family,
3370 or the *struct* types of the Lisp family.
3372 New instances of a `struct` can be constructed with a [struct
3373 expression](#struct-expressions).
3375 The memory layout of a `struct` is undefined by default to allow for compiler
3376 optimizations like field reordering, but it can be fixed with the
3377 `#[repr(...)]` attribute. In either case, fields may be given in any order in
3378 a corresponding struct *expression*; the resulting `struct` value will always
3379 have the same memory layout.
3381 The fields of a `struct` may be qualified by [visibility
3382 modifiers](#visibility-and-privacy), to allow access to data in a
3383 struct outside a module.
3385 A _tuple struct_ type is just like a struct type, except that the fields are
3388 A _unit-like struct_ type is like a struct type, except that it has no
3389 fields. The one value constructed by the associated [struct
3390 expression](#struct-expressions) is the only value that inhabits such a
3393 ### Enumerated types
3395 An *enumerated type* is a nominal, heterogeneous disjoint union type, denoted
3396 by the name of an [`enum` item](#enumerations). [^enumtype]
3398 [^enumtype]: The `enum` type is analogous to a `data` constructor declaration in
3399 ML, or a *pick ADT* in Limbo.
3401 An [`enum` item](#enumerations) declares both the type and a number of *variant
3402 constructors*, each of which is independently named and takes an optional tuple
3405 New instances of an `enum` can be constructed by calling one of the variant
3406 constructors, in a [call expression](#call-expressions).
3408 Any `enum` value consumes as much memory as the largest variant constructor for
3409 its corresponding `enum` type.
3411 Enum types cannot be denoted *structurally* as types, but must be denoted by
3412 named reference to an [`enum` item](#enumerations).
3416 Nominal types — [enumerations](#enumerated-types) and
3417 [structs](#struct-types) — may be recursive. That is, each `enum`
3418 constructor or `struct` field may refer, directly or indirectly, to the
3419 enclosing `enum` or `struct` type itself. Such recursion has restrictions:
3421 * Recursive types must include a nominal type in the recursion
3422 (not mere [type definitions](grammar.html#type-definitions),
3423 or other structural types such as [arrays](#array-and-slice-types) or [tuples](#tuple-types)).
3424 * A recursive `enum` item must have at least one non-recursive constructor
3425 (in order to give the recursion a basis case).
3426 * The size of a recursive type must be finite;
3427 in other words the recursive fields of the type must be [pointer types](#pointer-types).
3428 * Recursive type definitions can cross module boundaries, but not module *visibility* boundaries,
3429 or crate boundaries (in order to simplify the module system and type checker).
3431 An example of a *recursive* type and its use:
3436 Cons(T, Box<List<T>>)
3439 let a: List<i32> = List::Cons(7, Box::new(List::Cons(13, Box::new(List::Nil))));
3444 All pointers in Rust are explicit first-class values. They can be copied,
3445 stored into data structs, and returned from functions. There are two
3446 varieties of pointer in Rust:
3449 : These point to memory _owned by some other value_.
3450 A reference type is written `&type`,
3451 or `&'a type` when you need to specify an explicit lifetime.
3452 Copying a reference is a "shallow" operation:
3453 it involves only copying the pointer itself.
3454 Releasing a reference has no effect on the value it points to,
3455 but a reference of a temporary value will keep it alive during the scope
3456 of the reference itself.
3458 * Raw pointers (`*`)
3459 : Raw pointers are pointers without safety or liveness guarantees.
3460 Raw pointers are written as `*const T` or `*mut T`,
3461 for example `*const i32` means a raw pointer to a 32-bit integer.
3462 Copying or dropping a raw pointer has no effect on the lifecycle of any
3463 other value. Dereferencing a raw pointer or converting it to any other
3464 pointer type is an [`unsafe` operation](#unsafe-functions).
3465 Raw pointers are generally discouraged in Rust code;
3466 they exist to support interoperability with foreign code,
3467 and writing performance-critical or low-level functions.
3469 The standard library contains additional 'smart pointer' types beyond references
3474 The function type constructor `fn` forms new function types. A function type
3475 consists of a possibly-empty set of function-type modifiers (such as `unsafe`
3476 or `extern`), a sequence of input types and an output type.
3478 An example of a `fn` type:
3481 fn add(x: i32, y: i32) -> i32 {
3485 let mut x = add(5,7);
3487 type Binop = fn(i32, i32) -> i32;
3488 let bo: Binop = add;
3492 #### Function types for specific items
3494 Internal to the compiler, there are also function types that are specific to a particular
3495 function item. In the following snippet, for example, the internal types of the functions
3496 `foo` and `bar` are different, despite the fact that they have the same signature:
3503 The types of `foo` and `bar` can both be implicitly coerced to the fn
3504 pointer type `fn()`. There is currently no syntax for unique fn types,
3505 though the compiler will emit a type like `fn() {foo}` in error
3506 messages to indicate "the unique fn type for the function `foo`".
3510 A [lambda expression](#lambda-expressions) produces a closure value with
3511 a unique, anonymous type that cannot be written out.
3513 Depending on the requirements of the closure, its type implements one or
3514 more of the closure traits:
3517 : The closure can be called once. A closure called as `FnOnce`
3518 can move out values from its environment.
3521 : The closure can be called multiple times as mutable. A closure called as
3522 `FnMut` can mutate values from its environment. `FnMut` inherits from
3523 `FnOnce` (i.e. anything implementing `FnMut` also implements `FnOnce`).
3526 : The closure can be called multiple times through a shared reference.
3527 A closure called as `Fn` can neither move out from nor mutate values
3528 from its environment. `Fn` inherits from `FnMut`, which itself
3529 inherits from `FnOnce`.
3534 In Rust, a type like `&SomeTrait` or `Box<SomeTrait>` is called a _trait object_.
3535 Each instance of a trait object includes:
3537 - a pointer to an instance of a type `T` that implements `SomeTrait`
3538 - a _virtual method table_, often just called a _vtable_, which contains, for
3539 each method of `SomeTrait` that `T` implements, a pointer to `T`'s
3540 implementation (i.e. a function pointer).
3542 The purpose of trait objects is to permit "late binding" of methods. A call to
3543 a method on a trait object is only resolved to a vtable entry at compile time.
3544 The actual implementation for each vtable entry can vary on an object-by-object
3547 Note that for a trait object to be instantiated, the trait must be
3548 _object-safe_. Object safety rules are defined in [RFC 255].
3550 [RFC 255]: https://github.com/rust-lang/rfcs/blob/master/text/0255-object-safety.md
3552 Given a pointer-typed expression `E` of type `&T` or `Box<T>`, where `T`
3553 implements trait `R`, casting `E` to the corresponding pointer type `&R` or
3554 `Box<R>` results in a value of the _trait object_ `R`. This result is
3555 represented as a pair of pointers: the vtable pointer for the `T`
3556 implementation of `R`, and the pointer value of `E`.
3558 An example of a trait object:
3562 fn stringify(&self) -> String;
3565 impl Printable for i32 {
3566 fn stringify(&self) -> String { self.to_string() }
3569 fn print(a: Box<Printable>) {
3570 println!("{}", a.stringify());
3574 print(Box::new(10) as Box<Printable>);
3578 In this example, the trait `Printable` occurs as a trait object in both the
3579 type signature of `print`, and the cast expression in `main`.
3583 Within the body of an item that has type parameter declarations, the names of
3584 its type parameters are types:
3587 fn to_vec<A: Clone>(xs: &[A]) -> Vec<A> {
3591 let first: A = xs[0].clone();
3592 let mut rest: Vec<A> = to_vec(&xs[1..]);
3593 rest.insert(0, first);
3598 Here, `first` has type `A`, referring to `to_vec`'s `A` type parameter; and `rest`
3599 has type `Vec<A>`, a vector with element type `A`.
3603 The special type `Self` has a meaning within traits and impls. In a trait definition, it refers
3604 to an implicit type parameter representing the "implementing" type. In an impl,
3605 it is an alias for the implementing type. For example, in:
3609 fn make_string(&self) -> String;
3612 impl Printable for String {
3613 fn make_string(&self) -> String {
3619 The notation `&self` is a shorthand for `self: &Self`. In this case,
3620 in the impl, `Self` refers to the value of type `String` that is the
3621 receiver for a call to the method `make_string`.
3625 Subtyping is implicit and can occur at any stage in type checking or
3626 inference. Subtyping in Rust is very restricted and occurs only due to
3627 variance with respect to lifetimes and between types with higher ranked
3628 lifetimes. If we were to erase lifetimes from types, then the only subtyping
3629 would be due to type equality.
3631 Consider the following example: string literals always have `'static`
3632 lifetime. Nevertheless, we can assign `s` to `t`:
3636 let s: &'static str = "hi";
3640 Since `'static` "lives longer" than `'a`, `&'static str` is a subtype of
3645 Coercions are defined in [RFC401]. A coercion is implicit and has no syntax.
3647 [RFC401]: https://github.com/rust-lang/rfcs/blob/master/text/0401-coercions.md
3651 A coercion can only occur at certain coercion sites in a program; these are
3652 typically places where the desired type is explicit or can be derived by
3653 propagation from explicit types (without type inference). Possible coercion
3656 * `let` statements where an explicit type is given.
3658 For example, `128` is coerced to have type `i8` in the following:
3664 * `static` and `const` statements (similar to `let` statements).
3666 * Arguments for function calls
3668 The value being coerced is the actual parameter, and it is coerced to
3669 the type of the formal parameter.
3671 For example, `128` is coerced to have type `i8` in the following:
3681 * Instantiations of struct or variant fields
3683 For example, `128` is coerced to have type `i8` in the following:
3686 struct Foo { x: i8 }
3693 * Function results, either the final line of a block if it is not
3694 semicolon-terminated or any expression in a `return` statement
3696 For example, `128` is coerced to have type `i8` in the following:
3704 If the expression in one of these coercion sites is a coercion-propagating
3705 expression, then the relevant sub-expressions in that expression are also
3706 coercion sites. Propagation recurses from these new coercion sites.
3707 Propagating expressions and their relevant sub-expressions are:
3709 * Array literals, where the array has type `[U; n]`. Each sub-expression in
3710 the array literal is a coercion site for coercion to type `U`.
3712 * Array literals with repeating syntax, where the array has type `[U; n]`. The
3713 repeated sub-expression is a coercion site for coercion to type `U`.
3715 * Tuples, where a tuple is a coercion site to type `(U_0, U_1, ..., U_n)`.
3716 Each sub-expression is a coercion site to the respective type, e.g. the
3717 zeroth sub-expression is a coercion site to type `U_0`.
3719 * Parenthesized sub-expressions (`(e)`): if the expression has type `U`, then
3720 the sub-expression is a coercion site to `U`.
3722 * Blocks: if a block has type `U`, then the last expression in the block (if
3723 it is not semicolon-terminated) is a coercion site to `U`. This includes
3724 blocks which are part of control flow statements, such as `if`/`else`, if
3725 the block has a known type.
3729 Coercion is allowed between the following types:
3731 * `T` to `U` if `T` is a subtype of `U` (*reflexive case*)
3733 * `T_1` to `T_3` where `T_1` coerces to `T_2` and `T_2` coerces to `T_3`
3736 Note that this is not fully supported yet
3740 * `*mut T` to `*const T`
3742 * `&T` to `*const T`
3744 * `&mut T` to `*mut T`
3746 * `&T` to `&U` if `T` implements `Deref<Target = U>`. For example:
3749 use std::ops::Deref;
3751 struct CharContainer {
3755 impl Deref for CharContainer {
3758 fn deref<'a>(&'a self) -> &'a char {
3763 fn foo(arg: &char) {}
3766 let x = &mut CharContainer { value: 'y' };
3767 foo(x); //&mut CharContainer is coerced to &char.
3771 * `&mut T` to `&mut U` if `T` implements `DerefMut<Target = U>`.
3773 * TyCtor(`T`) to TyCtor(coerce_inner(`T`)), where TyCtor(`T`) is one of
3781 - coerce_inner(`[T, ..n]`) = `[T]`
3782 - coerce_inner(`T`) = `U` where `T` is a concrete type which implements the
3785 In the future, coerce_inner will be recursively extended to tuples and
3786 structs. In addition, coercions from sub-traits to super-traits will be
3787 added. See [RFC401] for more details.
3791 Several traits define special evaluation behavior.
3795 The `Copy` trait changes the semantics of a type implementing it. Values whose
3796 type implements `Copy` are copied rather than moved upon assignment.
3798 ## The `Sized` trait
3800 The `Sized` trait indicates that the size of this type is known at compile-time.
3804 The `Drop` trait provides a destructor, to be run whenever a value of this type
3807 ## The `Deref` trait
3809 The `Deref<Target = U>` trait allows a type to implicitly implement all the methods
3810 of the type `U`. When attempting to resolve a method call, the compiler will search
3811 the top-level type for the implementation of the called method. If no such method is
3812 found, `.deref()` is called and the compiler continues to search for the method
3813 implementation in the returned type `U`.
3817 A Rust program's memory consists of a static set of *items* and a *heap*.
3818 Immutable portions of the heap may be safely shared between threads, mutable
3819 portions may not be safely shared, but several mechanisms for effectively-safe
3820 sharing of mutable values, built on unsafe code but enforcing a safe locking
3821 discipline, exist in the standard library.
3823 Allocations in the stack consist of *variables*, and allocations in the heap
3826 ### Memory allocation and lifetime
3828 The _items_ of a program are those functions, modules and types that have their
3829 value calculated at compile-time and stored uniquely in the memory image of the
3830 rust process. Items are neither dynamically allocated nor freed.
3832 The _heap_ is a general term that describes boxes. The lifetime of an
3833 allocation in the heap depends on the lifetime of the box values pointing to
3834 it. Since box values may themselves be passed in and out of frames, or stored
3835 in the heap, heap allocations may outlive the frame they are allocated within.
3837 ### Memory ownership
3839 When a stack frame is exited, its local allocations are all released, and its
3840 references to boxes are dropped.
3844 A _variable_ is a component of a stack frame, either a named function parameter,
3845 an anonymous [temporary](#lvalues-rvalues-and-temporaries), or a named local
3848 A _local variable_ (or *stack-local* allocation) holds a value directly,
3849 allocated within the stack's memory. The value is a part of the stack frame.
3851 Local variables are immutable unless declared otherwise like: `let mut x = ...`.
3853 Function parameters are immutable unless declared with `mut`. The `mut` keyword
3854 applies only to the following parameter (so `|mut x, y|` and `fn f(mut x:
3855 Box<i32>, y: Box<i32>)` declare one mutable variable `x` and one immutable
3858 Methods that take either `self` or `Box<Self>` can optionally place them in a
3859 mutable variable by prefixing them with `mut` (similar to regular arguments):
3863 fn change(mut self) -> Self;
3864 fn modify(mut self: Box<Self>) -> Box<Self>;
3868 Local variables are not initialized when allocated; the entire frame worth of
3869 local variables are allocated at once, on frame-entry, in an uninitialized
3870 state. Subsequent statements within a function may or may not initialize the
3871 local variables. Local variables can be used only after they have been
3872 initialized; this is enforced by the compiler.
3876 The Rust compiler supports various methods to link crates together both
3877 statically and dynamically. This section will explore the various methods to
3878 link Rust crates together, and more information about native libraries can be
3879 found in the [FFI section of the book][ffi].
3881 In one session of compilation, the compiler can generate multiple artifacts
3882 through the usage of either command line flags or the `crate_type` attribute.
3883 If one or more command line flags are specified, all `crate_type` attributes will
3884 be ignored in favor of only building the artifacts specified by command line.
3886 * `--crate-type=bin`, `#[crate_type = "bin"]` - A runnable executable will be
3887 produced. This requires that there is a `main` function in the crate which
3888 will be run when the program begins executing. This will link in all Rust and
3889 native dependencies, producing a distributable binary.
3891 * `--crate-type=lib`, `#[crate_type = "lib"]` - A Rust library will be produced.
3892 This is an ambiguous concept as to what exactly is produced because a library
3893 can manifest itself in several forms. The purpose of this generic `lib` option
3894 is to generate the "compiler recommended" style of library. The output library
3895 will always be usable by rustc, but the actual type of library may change from
3896 time-to-time. The remaining output types are all different flavors of
3897 libraries, and the `lib` type can be seen as an alias for one of them (but the
3898 actual one is compiler-defined).
3900 * `--crate-type=dylib`, `#[crate_type = "dylib"]` - A dynamic Rust library will
3901 be produced. This is different from the `lib` output type in that this forces
3902 dynamic library generation. The resulting dynamic library can be used as a
3903 dependency for other libraries and/or executables. This output type will
3904 create `*.so` files on linux, `*.dylib` files on osx, and `*.dll` files on
3907 * `--crate-type=staticlib`, `#[crate_type = "staticlib"]` - A static system
3908 library will be produced. This is different from other library outputs in that
3909 the Rust compiler will never attempt to link to `staticlib` outputs. The
3910 purpose of this output type is to create a static library containing all of
3911 the local crate's code along with all upstream dependencies. The static
3912 library is actually a `*.a` archive on linux and osx and a `*.lib` file on
3913 windows. This format is recommended for use in situations such as linking
3914 Rust code into an existing non-Rust application because it will not have
3915 dynamic dependencies on other Rust code.
3917 * `--crate-type=rlib`, `#[crate_type = "rlib"]` - A "Rust library" file will be
3918 produced. This is used as an intermediate artifact and can be thought of as a
3919 "static Rust library". These `rlib` files, unlike `staticlib` files, are
3920 interpreted by the Rust compiler in future linkage. This essentially means
3921 that `rustc` will look for metadata in `rlib` files like it looks for metadata
3922 in dynamic libraries. This form of output is used to produce statically linked
3923 executables as well as `staticlib` outputs.
3925 Note that these outputs are stackable in the sense that if multiple are
3926 specified, then the compiler will produce each form of output at once without
3927 having to recompile. However, this only applies for outputs specified by the
3928 same method. If only `crate_type` attributes are specified, then they will all
3929 be built, but if one or more `--crate-type` command line flags are specified,
3930 then only those outputs will be built.
3932 With all these different kinds of outputs, if crate A depends on crate B, then
3933 the compiler could find B in various different forms throughout the system. The
3934 only forms looked for by the compiler, however, are the `rlib` format and the
3935 dynamic library format. With these two options for a dependent library, the
3936 compiler must at some point make a choice between these two formats. With this
3937 in mind, the compiler follows these rules when determining what format of
3938 dependencies will be used:
3940 1. If a static library is being produced, all upstream dependencies are
3941 required to be available in `rlib` formats. This requirement stems from the
3942 reason that a dynamic library cannot be converted into a static format.
3944 Note that it is impossible to link in native dynamic dependencies to a static
3945 library, and in this case warnings will be printed about all unlinked native
3946 dynamic dependencies.
3948 2. If an `rlib` file is being produced, then there are no restrictions on what
3949 format the upstream dependencies are available in. It is simply required that
3950 all upstream dependencies be available for reading metadata from.
3952 The reason for this is that `rlib` files do not contain any of their upstream
3953 dependencies. It wouldn't be very efficient for all `rlib` files to contain a
3954 copy of `libstd.rlib`!
3956 3. If an executable is being produced and the `-C prefer-dynamic` flag is not
3957 specified, then dependencies are first attempted to be found in the `rlib`
3958 format. If some dependencies are not available in an rlib format, then
3959 dynamic linking is attempted (see below).
3961 4. If a dynamic library or an executable that is being dynamically linked is
3962 being produced, then the compiler will attempt to reconcile the available
3963 dependencies in either the rlib or dylib format to create a final product.
3965 A major goal of the compiler is to ensure that a library never appears more
3966 than once in any artifact. For example, if dynamic libraries B and C were
3967 each statically linked to library A, then a crate could not link to B and C
3968 together because there would be two copies of A. The compiler allows mixing
3969 the rlib and dylib formats, but this restriction must be satisfied.
3971 The compiler currently implements no method of hinting what format a library
3972 should be linked with. When dynamically linking, the compiler will attempt to
3973 maximize dynamic dependencies while still allowing some dependencies to be
3974 linked in via an rlib.
3976 For most situations, having all libraries available as a dylib is recommended
3977 if dynamically linking. For other situations, the compiler will emit a
3978 warning if it is unable to determine which formats to link each library with.
3980 In general, `--crate-type=bin` or `--crate-type=lib` should be sufficient for
3981 all compilation needs, and the other options are just available if more
3982 fine-grained control is desired over the output format of a Rust crate.
3986 Unsafe operations are those that potentially violate the memory-safety
3987 guarantees of Rust's static semantics.
3989 The following language level features cannot be used in the safe subset of
3992 - Dereferencing a [raw pointer](#pointer-types).
3993 - Reading or writing a [mutable static variable](#mutable-statics).
3994 - Calling an unsafe function (including an intrinsic or foreign function).
3998 Unsafe functions are functions that are not safe in all contexts and/or for all
3999 possible inputs. Such a function must be prefixed with the keyword `unsafe` and
4000 can only be called from an `unsafe` block or another `unsafe` function.
4004 A block of code can be prefixed with the `unsafe` keyword, to permit calling
4005 `unsafe` functions or dereferencing raw pointers within a safe function.
4007 When a programmer has sufficient conviction that a sequence of potentially
4008 unsafe operations is actually safe, they can encapsulate that sequence (taken
4009 as a whole) within an `unsafe` block. The compiler will consider uses of such
4010 code safe, in the surrounding context.
4012 Unsafe blocks are used to wrap foreign libraries, make direct use of hardware
4013 or implement features not directly present in the language. For example, Rust
4014 provides the language features necessary to implement memory-safe concurrency
4015 in the language but the implementation of threads and message passing is in the
4018 Rust's type system is a conservative approximation of the dynamic safety
4019 requirements, so in some cases there is a performance cost to using safe code.
4020 For example, a doubly-linked list is not a tree structure and can only be
4021 represented with reference-counted pointers in safe code. By using `unsafe`
4022 blocks to represent the reverse links as raw pointers, it can be implemented
4025 ## Behavior considered undefined
4027 The following is a list of behavior which is forbidden in all Rust code,
4028 including within `unsafe` blocks and `unsafe` functions. Type checking provides
4029 the guarantee that these issues are never caused by safe code.
4032 * Dereferencing a null/dangling raw pointer
4033 * Reads of [undef](http://llvm.org/docs/LangRef.html#undefined-values)
4034 (uninitialized) memory
4035 * Breaking the [pointer aliasing
4036 rules](http://llvm.org/docs/LangRef.html#pointer-aliasing-rules)
4037 with raw pointers (a subset of the rules used by C)
4038 * `&mut` and `&` follow LLVM’s scoped [noalias] model, except if the `&T`
4039 contains an `UnsafeCell<U>`. Unsafe code must not violate these aliasing
4041 * Mutating non-mutable data (that is, data reached through a shared reference or
4042 data owned by a `let` binding), unless that data is contained within an `UnsafeCell<U>`.
4043 * Invoking undefined behavior via compiler intrinsics:
4044 * Indexing outside of the bounds of an object with `std::ptr::offset`
4045 (`offset` intrinsic), with
4046 the exception of one byte past the end which is permitted.
4047 * Using `std::ptr::copy_nonoverlapping_memory` (`memcpy32`/`memcpy64`
4048 intrinsics) on overlapping buffers
4049 * Invalid values in primitive types, even in private fields/locals:
4050 * Dangling/null references or boxes
4051 * A value other than `false` (0) or `true` (1) in a `bool`
4052 * A discriminant in an `enum` not included in the type definition
4053 * A value in a `char` which is a surrogate or above `char::MAX`
4054 * Non-UTF-8 byte sequences in a `str`
4055 * Unwinding into Rust from foreign code or unwinding from Rust into foreign
4056 code. Rust's failure system is not compatible with exception handling in
4057 other languages. Unwinding must be caught and handled at FFI boundaries.
4059 [noalias]: http://llvm.org/docs/LangRef.html#noalias
4061 ## Behavior not considered unsafe
4063 This is a list of behavior not considered *unsafe* in Rust terms, but that may
4067 * Leaks of memory and other resources
4068 * Exiting without calling destructors
4070 - Overflow is considered "unexpected" behavior and is always user-error,
4071 unless the `wrapping` primitives are used. In non-optimized builds, the compiler
4072 will insert debug checks that panic on overflow, but in optimized builds overflow
4073 instead results in wrapped values. See [RFC 560] for the rationale and more details.
4075 [RFC 560]: https://github.com/rust-lang/rfcs/blob/master/text/0560-integer-overflow.md
4077 # Appendix: Influences
4079 Rust is not a particularly original language, with design elements coming from
4080 a wide range of sources. Some of these are listed below (including elements
4081 that have since been removed):
4083 * SML, OCaml: algebraic data types, pattern matching, type inference,
4084 semicolon statement separation
4085 * C++: references, RAII, smart pointers, move semantics, monomorphization,
4087 * ML Kit, Cyclone: region based memory management
4088 * Haskell (GHC): typeclasses, type families
4089 * Newsqueak, Alef, Limbo: channels, concurrency
4090 * Erlang: message passing, thread failure, ~~linked thread failure~~,
4091 ~~lightweight concurrency~~
4092 * Swift: optional bindings
4093 * Scheme: hygienic macros
4095 * Ruby: ~~block syntax~~
4096 * NIL, Hermes: ~~typestate~~
4097 * [Unicode Annex #31](http://www.unicode.org/reports/tr31/): identifier and
4100 [ffi]: book/ffi.html
4101 [plugin]: book/compiler-plugins.html