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.
79 - The first character has property `XID_start`
80 - The remaining characters have property `XID_continue`
82 that does _not_ occur in the set of [keywords][keywords].
84 > **Note**: `XID_start` and `XID_continue` as character properties cover the
85 > character ranges used to form the more familiar C and Java language-family
90 Comments in Rust code follow the general C++ style of line (`//`) and
91 block (`/* ... */`) comment forms. Nested block comments are supported.
93 Line comments beginning with exactly _three_ slashes (`///`), and block
94 comments beginning with exactly one repeated asterisk in the block-open
95 sequence (`/**`), are interpreted as a special syntax for `doc`
96 [attributes](#attributes). That is, they are equivalent to writing
97 `#[doc="..."]` around the body of the comment, i.e., `/// Foo` turns into
100 Line comments beginning with `//!` and block comments beginning with `/*!` are
101 doc comments that apply to the parent of the comment, rather than the item
102 that follows. That is, they are equivalent to writing `#![doc="..."]` around
103 the body of the comment. `//!` comments are usually used to document
104 modules that occupy a source file.
106 Non-doc comments are interpreted as a form of whitespace.
110 Whitespace is any non-empty string containing only the following characters:
112 - `U+0020` (space, `' '`)
113 - `U+0009` (tab, `'\t'`)
114 - `U+000A` (LF, `'\n'`)
115 - `U+000D` (CR, `'\r'`)
117 Rust is a "free-form" language, meaning that all forms of whitespace serve only
118 to separate _tokens_ in the grammar, and have no semantic significance.
120 A Rust program has identical meaning if each whitespace element is replaced
121 with any other legal whitespace element, such as a single space character.
125 Tokens are primitive productions in the grammar defined by regular
126 (non-recursive) languages. "Simple" tokens are given in [string table
127 production](#string-table-productions) form, and occur in the rest of the
128 grammar as double-quoted strings. Other tokens have exact rules given.
132 A literal is an expression consisting of a single token, rather than a sequence
133 of tokens, that immediately and directly denotes the value it evaluates to,
134 rather than referring to it by name or some other evaluation rule. A literal is
135 a form of constant expression, so is evaluated (primarily) at compile time.
139 ##### Characters and strings
141 | | Example | `#` sets | Characters | Escapes |
142 |----------------------------------------------|-----------------|------------|-------------|---------------------|
143 | [Character](#character-literals) | `'H'` | `N/A` | All Unicode | `\'` & [Byte](#byte-escapes) & [Unicode](#unicode-escapes) |
144 | [String](#string-literals) | `"hello"` | `N/A` | All Unicode | `\"` & [Byte](#byte-escapes) & [Unicode](#unicode-escapes) |
145 | [Raw](#raw-string-literals) | `r#"hello"#` | `0...` | All Unicode | `N/A` |
146 | [Byte](#byte-literals) | `b'H'` | `N/A` | All ASCII | `\'` & [Byte](#byte-escapes) |
147 | [Byte string](#byte-string-literals) | `b"hello"` | `N/A` | All ASCII | `\"` & [Byte](#byte-escapes) |
148 | [Raw byte string](#raw-byte-string-literals) | `br#"hello"#` | `0...` | All ASCII | `N/A` |
154 | `\x7F` | 8-bit character code (exactly 2 digits) |
156 | `\r` | Carriage return |
160 ##### Unicode escapes
163 | `\u{7FFF}` | 24-bit Unicode character code (up to 6 digits) |
167 | [Number literals](#number-literals)`*` | Example | Exponentiation | Suffixes |
168 |----------------------------------------|---------|----------------|----------|
169 | Decimal integer | `98_222` | `N/A` | Integer suffixes |
170 | Hex integer | `0xff` | `N/A` | Integer suffixes |
171 | Octal integer | `0o77` | `N/A` | Integer suffixes |
172 | Binary integer | `0b1111_0000` | `N/A` | Integer suffixes |
173 | Floating-point | `123.0E+77` | `Optional` | Floating-point suffixes |
175 `*` All number literals allow `_` as a visual separator: `1_234.0E+18f64`
178 | Integer | Floating-point |
179 |---------|----------------|
180 | `u8`, `i8`, `u16`, `i16`, `u32`, `i32`, `u64`, `i64`, `isize`, `usize` | `f32`, `f64` |
182 #### Character and string literals
184 ##### Character literals
186 A _character literal_ is a single Unicode character enclosed within two
187 `U+0027` (single-quote) characters, with the exception of `U+0027` itself,
188 which must be _escaped_ by a preceding `U+005C` character (`\`).
190 ##### String literals
192 A _string literal_ is a sequence of any Unicode characters enclosed within two
193 `U+0022` (double-quote) characters, with the exception of `U+0022` itself,
194 which must be _escaped_ by a preceding `U+005C` character (`\`).
196 Line-break characters are allowed in string literals. Normally they represent
197 themselves (i.e. no translation), but as a special exception, when a `U+005C`
198 character (`\`) occurs immediately before the newline, the `U+005C` character,
199 the newline, and all whitespace at the beginning of the next line are ignored.
200 Thus `a` and `b` are equal:
210 ##### Character escapes
212 Some additional _escapes_ are available in either character or non-raw string
213 literals. An escape starts with a `U+005C` (`\`) and continues with one of the
216 * An _8-bit code point escape_ starts with `U+0078` (`x`) and is
217 followed by exactly two _hex digits_. It denotes the Unicode code point
218 equal to the provided hex value.
219 * A _24-bit code point escape_ starts with `U+0075` (`u`) and is followed
220 by up to six _hex digits_ surrounded by braces `U+007B` (`{`) and `U+007D`
221 (`}`). It denotes the Unicode code point equal to the provided hex value.
222 * A _whitespace escape_ is one of the characters `U+006E` (`n`), `U+0072`
223 (`r`), or `U+0074` (`t`), denoting the Unicode values `U+000A` (LF),
224 `U+000D` (CR) or `U+0009` (HT) respectively.
225 * The _backslash escape_ is the character `U+005C` (`\`) which must be
226 escaped in order to denote *itself*.
228 ##### Raw string literals
230 Raw string literals do not process any escapes. They start with the character
231 `U+0072` (`r`), followed by zero or more of the character `U+0023` (`#`) and a
232 `U+0022` (double-quote) character. The _raw string body_ can contain any sequence
233 of Unicode characters and is terminated only by another `U+0022` (double-quote)
234 character, followed by the same number of `U+0023` (`#`) characters that preceded
235 the opening `U+0022` (double-quote) character.
237 All Unicode characters contained in the raw string body represent themselves,
238 the characters `U+0022` (double-quote) (except when followed by at least as
239 many `U+0023` (`#`) characters as were used to start the raw string literal) or
240 `U+005C` (`\`) do not have any special meaning.
242 Examples for string literals:
245 "foo"; r"foo"; // foo
246 "\"foo\""; r#""foo""#; // "foo"
249 r##"foo #"# bar"##; // foo #"# bar
251 "\x52"; "R"; r"R"; // R
252 "\\x52"; r"\x52"; // \x52
255 #### Byte and byte string literals
259 A _byte literal_ is a single ASCII character (in the `U+0000` to `U+007F`
260 range) or a single _escape_ preceded by the characters `U+0062` (`b`) and
261 `U+0027` (single-quote), and followed by the character `U+0027`. If the character
262 `U+0027` is present within the literal, it must be _escaped_ by a preceding
263 `U+005C` (`\`) character. It is equivalent to a `u8` unsigned 8-bit integer
266 ##### Byte string literals
268 A non-raw _byte string literal_ is a sequence of ASCII characters and _escapes_,
269 preceded by the characters `U+0062` (`b`) and `U+0022` (double-quote), and
270 followed by the character `U+0022`. If the character `U+0022` is present within
271 the literal, it must be _escaped_ by a preceding `U+005C` (`\`) character.
272 Alternatively, a byte string literal can be a _raw byte string literal_, defined
273 below. A byte string literal of length `n` is equivalent to a `&'static [u8; n]` borrowed fixed-sized array
274 of unsigned 8-bit integers.
276 Some additional _escapes_ are available in either byte or non-raw byte string
277 literals. An escape starts with a `U+005C` (`\`) and continues with one of the
280 * A _byte escape_ escape starts with `U+0078` (`x`) and is
281 followed by exactly two _hex digits_. It denotes the byte
282 equal to the provided hex value.
283 * A _whitespace escape_ is one of the characters `U+006E` (`n`), `U+0072`
284 (`r`), or `U+0074` (`t`), denoting the bytes values `0x0A` (ASCII LF),
285 `0x0D` (ASCII CR) or `0x09` (ASCII HT) respectively.
286 * The _backslash escape_ is the character `U+005C` (`\`) which must be
287 escaped in order to denote its ASCII encoding `0x5C`.
289 ##### Raw byte string literals
291 Raw byte string literals do not process any escapes. They start with the
292 character `U+0062` (`b`), followed by `U+0072` (`r`), followed by zero or more
293 of the character `U+0023` (`#`), and a `U+0022` (double-quote) character. The
294 _raw string body_ can contain any sequence of ASCII characters and is terminated
295 only by another `U+0022` (double-quote) character, followed by the same number of
296 `U+0023` (`#`) characters that preceded the opening `U+0022` (double-quote)
297 character. A raw byte string literal can not contain any non-ASCII byte.
299 All characters contained in the raw string body represent their ASCII encoding,
300 the characters `U+0022` (double-quote) (except when followed by at least as
301 many `U+0023` (`#`) characters as were used to start the raw string literal) or
302 `U+005C` (`\`) do not have any special meaning.
304 Examples for byte string literals:
307 b"foo"; br"foo"; // foo
308 b"\"foo\""; br#""foo""#; // "foo"
311 br##"foo #"# bar"##; // foo #"# bar
313 b"\x52"; b"R"; br"R"; // R
314 b"\\x52"; br"\x52"; // \x52
319 A _number literal_ is either an _integer literal_ or a _floating-point
320 literal_. The grammar for recognizing the two kinds of literals is mixed.
322 ##### Integer literals
324 An _integer literal_ has one of four forms:
326 * A _decimal literal_ starts with a *decimal digit* and continues with any
327 mixture of *decimal digits* and _underscores_.
328 * A _hex literal_ starts with the character sequence `U+0030` `U+0078`
329 (`0x`) and continues as any mixture of hex digits and underscores.
330 * An _octal literal_ starts with the character sequence `U+0030` `U+006F`
331 (`0o`) and continues as any mixture of octal digits and underscores.
332 * A _binary literal_ starts with the character sequence `U+0030` `U+0062`
333 (`0b`) and continues as any mixture of binary digits and underscores.
335 Like any literal, an integer literal may be followed (immediately,
336 without any spaces) by an _integer suffix_, which forcibly sets the
337 type of the literal. The integer suffix must be the name of one of the
338 integral types: `u8`, `i8`, `u16`, `i16`, `u32`, `i32`, `u64`, `i64`,
341 The type of an _unsuffixed_ integer literal is determined by type inference.
342 If an integer type can be _uniquely_ determined from the surrounding program
343 context, the unsuffixed integer literal has that type. If the program context
344 underconstrains the type, it defaults to the signed 32-bit integer `i32`; if
345 the program context overconstrains the type, it is considered a static type
348 Examples of integer literals of various forms:
355 0o70_i16; // type i16
356 0b1111_1111_1001_0000_i32; // type i32
357 0usize; // type usize
360 ##### Floating-point literals
362 A _floating-point literal_ has one of two forms:
364 * A _decimal literal_ followed by a period character `U+002E` (`.`). This is
365 optionally followed by another decimal literal, with an optional _exponent_.
366 * A single _decimal literal_ followed by an _exponent_.
368 Like integer literals, a floating-point literal may be followed by a
369 suffix, so long as the pre-suffix part does not end with `U+002E` (`.`).
370 The suffix forcibly sets the type of the literal. There are two valid
371 _floating-point suffixes_, `f32` and `f64` (the 32-bit and 64-bit floating point
372 types), which explicitly determine the type of the literal.
374 The type of an _unsuffixed_ floating-point literal is determined by type
375 inference. If a floating-point type can be _uniquely_ determined from the
376 surrounding program context, the unsuffixed floating-point literal has that type.
377 If the program context underconstrains the type, it defaults to double-precision `f64`;
378 if the program context overconstrains the type, it is considered a static type
381 Examples of floating-point literals of various forms:
384 123.0f64; // type f64
387 12E+99_f64; // type f64
388 let x: f64 = 2.; // type f64
391 This last example is different because it is not possible to use the suffix
392 syntax with a floating point literal ending in a period. `2.f64` would attempt
393 to call a method named `f64` on `2`.
395 The representation semantics of floating-point numbers are described in
396 ["Machine Types"](#machine-types).
398 #### Boolean literals
400 The two values of the boolean type are written `true` and `false`.
404 Symbols are a general class of printable [token](#tokens) that play structural
405 roles in a variety of grammar productions. They are catalogued here for
406 completeness as the set of remaining miscellaneous printable tokens that do not
407 otherwise appear as [unary operators](#unary-operator-expressions), [binary
408 operators](#binary-operator-expressions), or [keywords][keywords].
413 A _path_ is a sequence of one or more path components _logically_ separated by
414 a namespace qualifier (`::`). If a path consists of only one component, it may
415 refer to either an [item](#items) or a [variable](#variables) in a local control
416 scope. If a path has multiple components, it refers to an item.
418 Every item has a _canonical path_ within its crate, but the path naming an item
419 is only meaningful within a given crate. There is no global namespace across
420 crates; an item's canonical path merely identifies it within the crate.
422 Two examples of simple paths consisting of only identifier components:
429 Path components are usually [identifiers](#identifiers), but they may
430 also include angle-bracket-enclosed lists of type arguments. In
431 [expression](#expressions) context, the type argument list is given
432 after a `::` namespace qualifier in order to disambiguate it from a
433 relational expression involving the less-than symbol (`<`). In type
434 expression context, the final namespace qualifier is omitted.
436 Two examples of paths with type arguments:
439 # struct HashMap<K, V>(K,V);
441 # fn id<T>(t: T) -> T { t }
442 type T = HashMap<i32,String>; // Type arguments used in a type expression
443 let x = id::<i32>(10); // Type arguments used in a call expression
447 Paths can be denoted with various leading qualifiers to change the meaning of
450 * Paths starting with `::` are considered to be global paths where the
451 components of the path start being resolved from the crate root. Each
452 identifier in the path must resolve to an item.
460 ::a::foo(); // call a's foo function
466 * Paths starting with the keyword `super` begin resolution relative to the
467 parent module. Each further identifier must resolve to an item.
475 super::a::foo(); // call a's foo function
481 * Paths starting with the keyword `self` begin resolution relative to the
482 current module. Each further identifier must resolve to an item.
494 A number of minor features of Rust are not central enough to have their own
495 syntax, and yet are not implementable as functions. Instead, they are given
496 names, and invoked through a consistent syntax: `some_extension!(...)`.
498 Users of `rustc` can define new syntax extensions in two ways:
500 * [Compiler plugins][plugin] can include arbitrary Rust code that
501 manipulates syntax trees at compile time. Note that the interface
502 for compiler plugins is considered highly unstable.
504 * [Macros](book/macros.html) define new syntax in a higher-level,
509 `macro_rules` allows users to define syntax extension in a declarative way. We
510 call such extensions "macros by example" or simply "macros" — to be distinguished
511 from the "procedural macros" defined in [compiler plugins][plugin].
513 Currently, macros can expand to expressions, statements, items, or patterns.
515 (A `sep_token` is any token other than `*` and `+`. A `non_special_token` is
516 any token other than a delimiter or `$`.)
518 The macro expander looks up macro invocations by name, and tries each macro
519 rule in turn. It transcribes the first successful match. Matching and
520 transcription are closely related to each other, and we will describe them
525 The macro expander matches and transcribes every token that does not begin with
526 a `$` literally, including delimiters. For parsing reasons, delimiters must be
527 balanced, but they are otherwise not special.
529 In the matcher, `$` _name_ `:` _designator_ matches the nonterminal in the Rust
530 syntax named by _designator_. Valid designators are `item`, `block`, `stmt`,
531 `pat`, `expr`, `ty` (type), `ident`, `path`, `tt` (either side of the `=>`
532 in macro rules). In the transcriber, the designator is already known, and so
533 only the name of a matched nonterminal comes after the dollar sign.
535 In both the matcher and transcriber, the Kleene star-like operator indicates
536 repetition. The Kleene star operator consists of `$` and parentheses, optionally
537 followed by a separator token, followed by `*` or `+`. `*` means zero or more
538 repetitions, `+` means at least one repetition. The parentheses are not matched or
539 transcribed. On the matcher side, a name is bound to _all_ of the names it
540 matches, in a structure that mimics the structure of the repetition encountered
541 on a successful match. The job of the transcriber is to sort that structure
544 The rules for transcription of these repetitions are called "Macro By Example".
545 Essentially, one "layer" of repetition is discharged at a time, and all of them
546 must be discharged by the time a name is transcribed. Therefore, `( $( $i:ident
547 ),* ) => ( $i )` is an invalid macro, but `( $( $i:ident ),* ) => ( $( $i:ident
548 ),* )` is acceptable (if trivial).
550 When Macro By Example encounters a repetition, it examines all of the `$`
551 _name_ s that occur in its body. At the "current layer", they all must repeat
552 the same number of times, so ` ( $( $i:ident ),* ; $( $j:ident ),* ) => ( $(
553 ($i,$j) ),* )` is valid if given the argument `(a,b,c ; d,e,f)`, but not
554 `(a,b,c ; d,e)`. The repetition walks through the choices at that layer in
555 lockstep, so the former input transcribes to `(a,d), (b,e), (c,f)`.
557 Nested repetitions are allowed.
559 ### Parsing limitations
561 The parser used by the macro system is reasonably powerful, but the parsing of
562 Rust syntax is restricted in two ways:
564 1. Macro definitions are required to include suitable separators after parsing
565 expressions and other bits of the Rust grammar. This implies that
566 a macro definition like `$i:expr [ , ]` is not legal, because `[` could be part
567 of an expression. A macro definition like `$i:expr,` or `$i:expr;` would be legal,
568 however, because `,` and `;` are legal separators. See [RFC 550] for more information.
569 2. The parser must have eliminated all ambiguity by the time it reaches a `$`
570 _name_ `:` _designator_. This requirement most often affects name-designator
571 pairs when they occur at the beginning of, or immediately after, a `$(...)*`;
572 requiring a distinctive token in front can solve the problem.
574 [RFC 550]: https://github.com/rust-lang/rfcs/blob/master/text/0550-macro-future-proofing.md
576 # Crates and source files
578 Although Rust, like any other language, can be implemented by an interpreter as
579 well as a compiler, the only existing implementation is a compiler —
580 from now on referred to as *the* Rust compiler — and the language has
581 always been designed to be compiled. For these reasons, this section assumes a
584 Rust's semantics obey a *phase distinction* between compile-time and
585 run-time.[^phase-distinction] Those semantic rules that have a *static
586 interpretation* govern the success or failure of compilation. Those semantics
587 that have a *dynamic interpretation* govern the behavior of the program at
590 [^phase-distinction]: This distinction would also exist in an interpreter.
591 Static checks like syntactic analysis, type checking, and lints should
592 happen before the program is executed regardless of when it is executed.
594 The compilation model centers on artifacts called _crates_. Each compilation
595 processes a single crate in source form, and if successful, produces a single
596 crate in binary form: either an executable or some sort of
597 library.[^cratesourcefile]
599 [^cratesourcefile]: A crate is somewhat analogous to an *assembly* in the
600 ECMA-335 CLI model, a *library* in the SML/NJ Compilation Manager, a *unit*
601 in the Owens and Flatt module system, or a *configuration* in Mesa.
603 A _crate_ is a unit of compilation and linking, as well as versioning,
604 distribution and runtime loading. A crate contains a _tree_ of nested
605 [module](#modules) scopes. The top level of this tree is a module that is
606 anonymous (from the point of view of paths within the module) and any item
607 within a crate has a canonical [module path](#paths) denoting its location
608 within the crate's module tree.
610 The Rust compiler is always invoked with a single source file as input, and
611 always produces a single output crate. The processing of that source file may
612 result in other source files being loaded as modules. Source files have the
615 A Rust source file describes a module, the name and location of which —
616 in the module tree of the current crate — are defined from outside the
617 source file: either by an explicit `mod_item` in a referencing source file, or
618 by the name of the crate itself. Every source file is a module, but not every
619 module needs its own source file: [module definitions](#modules) can be nested
622 Each source file contains a sequence of zero or more `item` definitions, and
623 may optionally begin with any number of [attributes](#items-and-attributes)
624 that apply to the containing module, most of which influence the behavior of
625 the compiler. The anonymous crate module can have additional attributes that
626 apply to the crate as a whole.
629 // Specify the crate name.
630 #![crate_name = "projx"]
632 // Specify the type of output artifact.
633 #![crate_type = "lib"]
635 // Turn on a warning.
636 // This can be done in any module, not just the anonymous crate module.
637 #![warn(non_camel_case_types)]
640 A crate that contains a `main` function can be compiled to an executable. If a
641 `main` function is present, its return type must be [`unit`](#tuple-types)
642 and it must take no arguments.
644 # Items and attributes
646 Crates contain [items](#items), each of which may have some number of
647 [attributes](#attributes) attached to it.
651 An _item_ is a component of a crate. Items are organized within a crate by a
652 nested set of [modules](#modules). Every crate has a single "outermost"
653 anonymous module; all further items within the crate have [paths](#paths)
654 within the module tree of the crate.
656 Items are entirely determined at compile-time, generally remain fixed during
657 execution, and may reside in read-only memory.
659 There are several kinds of item:
661 * [`extern crate` declarations](#extern-crate-declarations)
662 * [`use` declarations](#use-declarations)
663 * [modules](#modules)
664 * [functions](#functions)
665 * [type definitions](grammar.html#type-definitions)
666 * [structures](#structures)
667 * [enumerations](#enumerations)
668 * [constant items](#constant-items)
669 * [static items](#static-items)
671 * [implementations](#implementations)
673 Some items form an implicit scope for the declaration of sub-items. In other
674 words, within a function or module, declarations of items can (in many cases)
675 be mixed with the statements, control blocks, and similar artifacts that
676 otherwise compose the item body. The meaning of these scoped items is the same
677 as if the item was declared outside the scope — it is still a static item
678 — except that the item's *path name* within the module namespace is
679 qualified by the name of the enclosing item, or is private to the enclosing
680 item (in the case of functions). The grammar specifies the exact locations in
681 which sub-item declarations may appear.
685 All items except modules, constants and statics may be *parameterized* by type.
686 Type parameters are given as a comma-separated list of identifiers enclosed in
687 angle brackets (`<...>`), after the name of the item and before its definition.
688 The type parameters of an item are considered "part of the name", not part of
689 the type of the item. A referencing [path](#paths) must (in principle) provide
690 type arguments as a list of comma-separated types enclosed within angle
691 brackets, in order to refer to the type-parameterized item. In practice, the
692 type-inference system can usually infer such argument types from context. There
693 are no general type-parametric types, only type-parametric items. That is, Rust
694 has no notion of type abstraction: there are no higher-ranked (or "forall") types
695 abstracted over other types, though higher-ranked types do exist for lifetimes.
699 A module is a container for zero or more [items](#items).
701 A _module item_ is a module, surrounded in braces, named, and prefixed with the
702 keyword `mod`. A module item introduces a new, named module into the tree of
703 modules making up a crate. Modules can nest arbitrarily.
705 An example of a module:
709 type Complex = (f64, f64);
710 fn sin(f: f64) -> f64 {
714 fn cos(f: f64) -> f64 {
718 fn tan(f: f64) -> f64 {
725 Modules and types share the same namespace. Declaring a named type with
726 the same name as a module in scope is forbidden: that is, a type definition,
727 trait, struct, enumeration, or type parameter can't shadow the name of a module
728 in scope, or vice versa.
730 A module without a body is loaded from an external file, by default with the
731 same name as the module, plus the `.rs` extension. When a nested submodule is
732 loaded from an external file, it is loaded from a subdirectory path that
733 mirrors the module hierarchy.
736 // Load the `vec` module from `vec.rs`
740 // Load the `local_data` module from `thread/local_data.rs`
741 // or `thread/local_data/mod.rs`.
746 The directories and files used for loading external file modules can be
747 influenced with the `path` attribute.
750 #[path = "thread_files"]
752 // Load the `local_data` module from `thread_files/tls.rs`
758 #### Extern crate declarations
760 An _`extern crate` declaration_ specifies a dependency on an external crate.
761 The external crate is then bound into the declaring scope as the `ident`
762 provided in the `extern_crate_decl`.
764 The external crate is resolved to a specific `soname` at compile time, and a
765 runtime linkage requirement to that `soname` is passed to the linker for
766 loading at runtime. The `soname` is resolved at compile time by scanning the
767 compiler's library path and matching the optional `crateid` provided against
768 the `crateid` attributes that were declared on the external crate when it was
769 compiled. If no `crateid` is provided, a default `name` attribute is assumed,
770 equal to the `ident` given in the `extern_crate_decl`.
772 Three examples of `extern crate` declarations:
777 extern crate std; // equivalent to: extern crate std as std;
779 extern crate std as ruststd; // linking to 'std' under another name
782 #### Use declarations
784 A _use declaration_ creates one or more local name bindings synonymous with
785 some other [path](#paths). Usually a `use` declaration is used to shorten the
786 path required to refer to a module item. These declarations may appear at the
787 top of [modules](#modules) and [blocks](grammar.html#block-expressions).
789 > **Note**: Unlike in many languages,
790 > `use` declarations in Rust do *not* declare linkage dependency with external crates.
791 > Rather, [`extern crate` declarations](#extern-crate-declarations) declare linkage dependencies.
793 Use declarations support a number of convenient shortcuts:
795 * Rebinding the target name as a new local name, using the syntax `use p::q::r as x;`
796 * Simultaneously binding a list of paths differing only in their final element,
797 using the glob-like brace syntax `use a::b::{c,d,e,f};`
798 * Binding all paths matching a given prefix, using the asterisk wildcard syntax
800 * Simultaneously binding a list of paths differing only in their final element
801 and their immediate parent module, using the `self` keyword, such as
802 `use a::b::{self, c, d};`
804 An example of `use` declarations:
807 use std::option::Option::{Some, None};
808 use std::collections::hash_map::{self, HashMap};
811 fn bar(map1: HashMap<String, usize>, map2: hash_map::HashMap<String, usize>){}
814 // Equivalent to 'foo(vec![std::option::Option::Some(1.0f64),
815 // std::option::Option::None]);'
816 foo(vec![Some(1.0f64), None]);
818 // Both `hash_map` and `HashMap` are in scope.
819 let map1 = HashMap::new();
820 let map2 = hash_map::HashMap::new();
825 Like items, `use` declarations are private to the containing module, by
826 default. Also like items, a `use` declaration can be public, if qualified by
827 the `pub` keyword. Such a `use` declaration serves to _re-export_ a name. A
828 public `use` declaration can therefore _redirect_ some public name to a
829 different target definition: even a definition with a private canonical path,
830 inside a different module. If a sequence of such redirections form a cycle or
831 cannot be resolved unambiguously, they represent a compile-time error.
833 An example of re-exporting:
838 pub use quux::foo::{bar, baz};
847 In this example, the module `quux` re-exports two public names defined in
850 Also note that the paths contained in `use` items are relative to the crate
851 root. So, in the previous example, the `use` refers to `quux::foo::{bar,
852 baz}`, and not simply to `foo::{bar, baz}`. This also means that top-level
853 module declarations should be at the crate root if direct usage of the declared
854 modules within `use` items is desired. It is also possible to use `self` and
855 `super` at the beginning of a `use` item to refer to the current and direct
856 parent modules respectively. All rules regarding accessing declared modules in
857 `use` declarations apply to both module declarations and `extern crate`
860 An example of what will and will not work for `use` items:
863 # #![allow(unused_imports)]
864 use foo::baz::foobaz; // good: foo is at the root of the crate
872 use foo::example::iter; // good: foo is at crate root
873 // use example::iter; // bad: core is not at the crate root
874 use self::baz::foobaz; // good: self refers to module 'foo'
875 use foo::bar::foobar; // good: foo is at crate root
882 use super::bar::foobar; // good: super refers to module 'foo'
892 A _function item_ defines a sequence of [statements](#statements) and an
893 optional final [expression](#expressions), along with a name and a set of
894 parameters. Functions are declared with the keyword `fn`. Functions declare a
895 set of *input* [*variables*](#variables) as parameters, through which the caller
896 passes arguments into the function, and the *output* [*type*](#types)
897 of the value the function will return to its caller on completion.
899 A function may also be copied into a first-class *value*, in which case the
900 value has the corresponding [*function type*](#function-types), and can be used
901 otherwise exactly as a function item (with a minor additional cost of calling
902 the function indirectly).
904 Every control path in a function logically ends with a `return` expression or a
905 diverging expression. If the outermost block of a function has a
906 value-producing expression in its final-expression position, that expression is
907 interpreted as an implicit `return` expression applied to the final-expression.
909 An example of a function:
912 fn add(x: i32, y: i32) -> i32 {
917 As with `let` bindings, function arguments are irrefutable patterns, so any
918 pattern that is valid in a let binding is also valid as an argument.
921 fn first((value, _): (i32, i32)) -> i32 { value }
925 #### Generic functions
927 A _generic function_ allows one or more _parameterized types_ to appear in its
928 signature. Each type parameter must be explicitly declared, in an
929 angle-bracket-enclosed, comma-separated list following the function name.
932 // foo is generic over A and B
934 fn foo<A, B>(x: A, y: B) {
937 Inside the function signature and body, the name of the type parameter can be
938 used as a type name. [Trait](#traits) bounds can be specified for type parameters
939 to allow methods with that trait to be called on values of that type. This is
940 specified using the `where` syntax:
943 fn foo<T>(x: T) where T: Debug {
946 When a generic function is referenced, its type is instantiated based on the
947 context of the reference. For example, calling the `foo` function here:
952 fn foo<T>(x: &[T]) where T: Debug {
960 will instantiate type parameter `T` with `i32`.
962 The type parameters can also be explicitly supplied in a trailing
963 [path](#paths) component after the function name. This might be necessary if
964 there is not sufficient context to determine the type parameters. For example,
965 `mem::size_of::<u32>() == 4`.
969 Unsafe operations are those that potentially violate the memory-safety
970 guarantees of Rust's static semantics.
972 The following language level features cannot be used in the safe subset of
975 - Dereferencing a [raw pointer](#pointer-types).
976 - Reading or writing a [mutable static variable](#mutable-statics).
977 - Calling an unsafe function (including an intrinsic or foreign function).
979 ##### Unsafe functions
981 Unsafe functions are functions that are not safe in all contexts and/or for all
982 possible inputs. Such a function must be prefixed with the keyword `unsafe` and
983 can only be called from an `unsafe` block or another `unsafe` function.
987 A block of code can be prefixed with the `unsafe` keyword, to permit calling
988 `unsafe` functions or dereferencing raw pointers within a safe function.
990 When a programmer has sufficient conviction that a sequence of potentially
991 unsafe operations is actually safe, they can encapsulate that sequence (taken
992 as a whole) within an `unsafe` block. The compiler will consider uses of such
993 code safe, in the surrounding context.
995 Unsafe blocks are used to wrap foreign libraries, make direct use of hardware
996 or implement features not directly present in the language. For example, Rust
997 provides the language features necessary to implement memory-safe concurrency
998 in the language but the implementation of threads and message passing is in the
1001 Rust's type system is a conservative approximation of the dynamic safety
1002 requirements, so in some cases there is a performance cost to using safe code.
1003 For example, a doubly-linked list is not a tree structure and can only be
1004 represented with reference-counted pointers in safe code. By using `unsafe`
1005 blocks to represent the reverse links as raw pointers, it can be implemented
1008 ##### Behavior considered undefined
1010 The following is a list of behavior which is forbidden in all Rust code,
1011 including within `unsafe` blocks and `unsafe` functions. Type checking provides
1012 the guarantee that these issues are never caused by safe code.
1015 * Dereferencing a null/dangling raw pointer
1016 * Reads of [undef](http://llvm.org/docs/LangRef.html#undefined-values)
1017 (uninitialized) memory
1018 * Breaking the [pointer aliasing
1019 rules](http://llvm.org/docs/LangRef.html#pointer-aliasing-rules)
1020 with raw pointers (a subset of the rules used by C)
1021 * `&mut` and `&` follow LLVM’s scoped [noalias] model, except if the `&T`
1022 contains an `UnsafeCell<U>`. Unsafe code must not violate these aliasing
1024 * Mutating non-mutable data (that is, data reached through a shared reference or
1025 data owned by a `let` binding), unless that data is contained within an `UnsafeCell<U>`.
1026 * Invoking undefined behavior via compiler intrinsics:
1027 * Indexing outside of the bounds of an object with `std::ptr::offset`
1028 (`offset` intrinsic), with
1029 the exception of one byte past the end which is permitted.
1030 * Using `std::ptr::copy_nonoverlapping_memory` (`memcpy32`/`memcpy64`
1031 intrinsics) on overlapping buffers
1032 * Invalid values in primitive types, even in private fields/locals:
1033 * Dangling/null references or boxes
1034 * A value other than `false` (0) or `true` (1) in a `bool`
1035 * A discriminant in an `enum` not included in the type definition
1036 * A value in a `char` which is a surrogate or above `char::MAX`
1037 * Non-UTF-8 byte sequences in a `str`
1038 * Unwinding into Rust from foreign code or unwinding from Rust into foreign
1039 code. Rust's failure system is not compatible with exception handling in
1040 other languages. Unwinding must be caught and handled at FFI boundaries.
1042 [noalias]: http://llvm.org/docs/LangRef.html#noalias
1044 ##### Behavior not considered unsafe
1046 This is a list of behavior not considered *unsafe* in Rust terms, but that may
1050 * Leaks of memory and other resources
1051 * Exiting without calling destructors
1053 - Overflow is considered "unexpected" behavior and is always user-error,
1054 unless the `wrapping` primitives are used. In non-optimized builds, the compiler
1055 will insert debug checks that panic on overflow, but in optimized builds overflow
1056 instead results in wrapped values. See [RFC 560] for the rationale and more details.
1058 [RFC 560]: https://github.com/rust-lang/rfcs/blob/master/text/0560-integer-overflow.md
1060 #### Diverging functions
1062 A special kind of function can be declared with a `!` character where the
1063 output type would normally be. For example:
1066 fn my_err(s: &str) -> ! {
1072 We call such functions "diverging" because they never return a value to the
1073 caller. Every control path in a diverging function must end with a `panic!()` or
1074 a call to another diverging function on every control path. The `!` annotation
1075 does *not* denote a type.
1077 It might be necessary to declare a diverging function because as mentioned
1078 previously, the typechecker checks that every control path in a function ends
1079 with a [`return`](#return-expressions) or diverging expression. So, if `my_err`
1080 were declared without the `!` annotation, the following code would not
1084 # fn my_err(s: &str) -> ! { panic!() }
1086 fn f(i: i32) -> i32 {
1091 my_err("Bad number!");
1096 This will not compile without the `!` annotation on `my_err`, since the `else`
1097 branch of the conditional in `f` does not return an `i32`, as required by the
1098 signature of `f`. Adding the `!` annotation to `my_err` informs the
1099 typechecker that, should control ever enter `my_err`, no further type judgments
1100 about `f` need to hold, since control will never resume in any context that
1101 relies on those judgments. Thus the return type on `f` only needs to reflect
1102 the `if` branch of the conditional.
1104 #### Extern functions
1106 Extern functions are part of Rust's foreign function interface, providing the
1107 opposite functionality to [external blocks](#external-blocks). Whereas
1108 external blocks allow Rust code to call foreign code, extern functions with
1109 bodies defined in Rust code _can be called by foreign code_. They are defined
1110 in the same way as any other Rust function, except that they have the `extern`
1114 // Declares an extern fn, the ABI defaults to "C"
1115 extern fn new_i32() -> i32 { 0 }
1117 // Declares an extern fn with "stdcall" ABI
1118 extern "stdcall" fn new_i32_stdcall() -> i32 { 0 }
1121 Unlike normal functions, extern fns have type `extern "ABI" fn()`. This is the
1122 same type as the functions declared in an extern block.
1125 # extern fn new_i32() -> i32 { 0 }
1126 let fptr: extern "C" fn() -> i32 = new_i32;
1129 Extern functions may be called directly from Rust code as Rust uses large,
1130 contiguous stack segments like C.
1134 A _type alias_ defines a new name for an existing [type](#types). Type
1135 aliases are declared with the keyword `type`. Every value has a single,
1136 specific type, but may implement several different traits, or be compatible with
1137 several different type constraints.
1139 For example, the following defines the type `Point` as a synonym for the type
1140 `(u8, u8)`, the type of pairs of unsigned 8 bit integers:
1143 type Point = (u8, u8);
1144 let p: Point = (41, 68);
1149 A _structure_ is a nominal [structure type](#structure-types) defined with the
1152 An example of a `struct` item and its use:
1155 struct Point {x: i32, y: i32}
1156 let p = Point {x: 10, y: 11};
1160 A _tuple structure_ is a nominal [tuple type](#tuple-types), also defined with
1161 the keyword `struct`. For example:
1164 struct Point(i32, i32);
1165 let p = Point(10, 11);
1166 let px: i32 = match p { Point(x, _) => x };
1169 A _unit-like struct_ is a structure without any fields, defined by leaving off
1170 the list of fields entirely. Such types will have a single value. For example:
1174 let c = [Cookie, Cookie, Cookie, Cookie];
1177 The precise memory layout of a structure is not specified. One can specify a
1178 particular layout using the [`repr` attribute](#ffi-attributes).
1182 An _enumeration_ is a simultaneous definition of a nominal [enumerated
1183 type](#enumerated-types) as well as a set of *constructors*, that can be used
1184 to create or pattern-match values of the corresponding enumerated type.
1186 Enumerations are declared with the keyword `enum`.
1188 An example of an `enum` item and its use:
1196 let mut a: Animal = Animal::Dog;
1200 Enumeration constructors can have either named or unnamed fields:
1205 Cat { name: String, weight: f64 }
1208 let mut a: Animal = Animal::Dog("Cocoa".to_string(), 37.2);
1209 a = Animal::Cat { name: "Spotty".to_string(), weight: 2.7 };
1212 In this example, `Cat` is a _struct-like enum variant_,
1213 whereas `Dog` is simply called an enum variant.
1215 Enums have a discriminant. You can assign them explicitly:
1223 If a discriminant isn't assigned, they start at zero, and add one for each
1226 You can cast an enum to get this value:
1229 # enum Foo { Bar = 123 }
1230 let x = Foo::Bar as u32; // x is now 123u32
1233 This only works as long as none of the variants have data attached. If
1234 it were `Bar(i32)`, this is disallowed.
1238 A *constant item* is a named _constant value_ which is not associated with a
1239 specific memory location in the program. Constants are essentially inlined
1240 wherever they are used, meaning that they are copied directly into the relevant
1241 context when used. References to the same constant are not necessarily
1242 guaranteed to refer to the same memory address.
1244 Constant values must not have destructors, and otherwise permit most forms of
1245 data. Constants may refer to the address of other constants, in which case the
1246 address will have the `static` lifetime. The compiler is, however, still at
1247 liberty to translate the constant many times, so the address referred to may not
1250 Constants must be explicitly typed. The type may be `bool`, `char`, a number, or
1251 a type derived from those primitive types. The derived types are references with
1252 the `static` lifetime, fixed-size arrays, tuples, enum variants, and structs.
1255 const BIT1: u32 = 1 << 0;
1256 const BIT2: u32 = 1 << 1;
1258 const BITS: [u32; 2] = [BIT1, BIT2];
1259 const STRING: &'static str = "bitstring";
1261 struct BitsNStrings<'a> {
1266 const BITS_N_STRINGS: BitsNStrings<'static> = BitsNStrings {
1274 A *static item* is similar to a *constant*, except that it represents a precise
1275 memory location in the program. A static is never "inlined" at the usage site,
1276 and all references to it refer to the same memory location. Static items have
1277 the `static` lifetime, which outlives all other lifetimes in a Rust program.
1278 Static items may be placed in read-only memory if they do not contain any
1279 interior mutability.
1281 Statics may contain interior mutability through the `UnsafeCell` language item.
1282 All access to a static is safe, but there are a number of restrictions on
1285 * Statics may not contain any destructors.
1286 * The types of static values must ascribe to `Sync` to allow threadsafe access.
1287 * Statics may not refer to other statics by value, only by reference.
1288 * Constants cannot refer to statics.
1290 Constants should in general be preferred over statics, unless large amounts of
1291 data are being stored, or single-address and mutability properties are required.
1293 #### Mutable statics
1295 If a static item is declared with the `mut` keyword, then it is allowed to
1296 be modified by the program. One of Rust's goals is to make concurrency bugs
1297 hard to run into, and this is obviously a very large source of race conditions
1298 or other bugs. For this reason, an `unsafe` block is required when either
1299 reading or writing a mutable static variable. Care should be taken to ensure
1300 that modifications to a mutable static are safe with respect to other threads
1301 running in the same process.
1303 Mutable statics are still very useful, however. They can be used with C
1304 libraries and can also be bound from C libraries (in an `extern` block).
1307 # fn atomic_add(_: &mut u32, _: u32) -> u32 { 2 }
1309 static mut LEVELS: u32 = 0;
1311 // This violates the idea of no shared state, and this doesn't internally
1312 // protect against races, so this function is `unsafe`
1313 unsafe fn bump_levels_unsafe1() -> u32 {
1319 // Assuming that we have an atomic_add function which returns the old value,
1320 // this function is "safe" but the meaning of the return value may not be what
1321 // callers expect, so it's still marked as `unsafe`
1322 unsafe fn bump_levels_unsafe2() -> u32 {
1323 return atomic_add(&mut LEVELS, 1);
1327 Mutable statics have the same restrictions as normal statics, except that the
1328 type of the value is not required to ascribe to `Sync`.
1332 A _trait_ describes an abstract interface that types can
1333 implement. This interface consists of associated items, which come in
1340 Associated functions whose first parameter is named `self` are called
1341 methods and may be invoked using `.` notation (e.g., `x.foo()`).
1343 All traits define an implicit type parameter `Self` that refers to
1344 "the type that is implementing this interface". Traits may also
1345 contain additional type parameters. These type parameters (including
1346 `Self`) may be constrained by other traits and so forth as usual.
1348 Trait bounds on `Self` are considered "supertraits". These are
1349 required to be acyclic. Supertraits are somewhat different from other
1350 constraints in that they affect what methods are available in the
1351 vtable when the trait is used as a [trait object](#trait-objects).
1353 Traits are implemented for specific types through separate
1354 [implementations](#implementations).
1356 Consider the following trait:
1359 # type Surface = i32;
1360 # type BoundingBox = i32;
1362 fn draw(&self, Surface);
1363 fn bounding_box(&self) -> BoundingBox;
1367 This defines a trait with two methods. All values that have
1368 [implementations](#implementations) of this trait in scope can have their
1369 `draw` and `bounding_box` methods called, using `value.bounding_box()`
1370 [syntax](#method-call-expressions).
1372 Traits can include default implementations of methods, as in:
1377 fn baz(&self) { println!("We called baz."); }
1381 Here the `baz` method has a default implementation, so types that implement
1382 `Foo` need only implement `bar`. It is also possible for implementing types
1383 to override a method that has a default implementation.
1385 Type parameters can be specified for a trait to make it generic. These appear
1386 after the trait name, using the same syntax used in [generic
1387 functions](#generic-functions).
1391 fn len(&self) -> u32;
1392 fn elt_at(&self, n: u32) -> T;
1393 fn iter<F>(&self, F) where F: Fn(T);
1397 It is also possible to define associated types for a trait. Consider the
1398 following example of a `Container` trait. Notice how the type is available
1399 for use in the method signatures:
1405 fn insert(&mut self, Self::E);
1409 In order for a type to implement this trait, it must not only provide
1410 implementations for every method, but it must specify the type `E`. Here's
1411 an implementation of `Container` for the standard library type `Vec`:
1416 # fn empty() -> Self;
1417 # fn insert(&mut self, Self::E);
1419 impl<T> Container for Vec<T> {
1421 fn empty() -> Vec<T> { Vec::new() }
1422 fn insert(&mut self, x: T) { self.push(x); }
1426 Generic functions may use traits as _bounds_ on their type parameters. This
1427 will have two effects:
1429 - Only types that have the trait may instantiate the parameter.
1430 - Within the generic function, the methods of the trait can be
1431 called on values that have the parameter's type.
1436 # type Surface = i32;
1437 # trait Shape { fn draw(&self, Surface); }
1438 fn draw_twice<T: Shape>(surface: Surface, sh: T) {
1444 Traits also define an [trait object](#trait-objects) with the same
1445 name as the trait. Values of this type are created by coercing from a
1446 pointer of some specific type to a pointer of trait type. For example,
1447 `&T` could be coerced to `&Shape` if `T: Shape` holds (and similarly
1448 for `Box<T>`). This coercion can either be implicit or
1449 [explicit](#type-cast-expressions). Here is an example of an explicit
1454 impl Shape for i32 { }
1455 let mycircle = 0i32;
1456 let myshape: Box<Shape> = Box::new(mycircle) as Box<Shape>;
1459 The resulting value is a box containing the value that was cast, along with
1460 information that identifies the methods of the implementation that was used.
1461 Values with a trait type can have [methods called](#method-call-expressions) on
1462 them, for any method in the trait, and can be used to instantiate type
1463 parameters that are bounded by the trait.
1465 Trait methods may be static, which means that they lack a `self` argument.
1466 This means that they can only be called with function call syntax (`f(x)`) and
1467 not method call syntax (`obj.f()`). The way to refer to the name of a static
1468 method is to qualify it with the trait name, treating the trait name like a
1469 module. For example:
1473 fn from_i32(n: i32) -> Self;
1476 fn from_i32(n: i32) -> f64 { n as f64 }
1478 let x: f64 = Num::from_i32(42);
1481 Traits may inherit from other traits. For example, in
1484 trait Shape { fn area(&self) -> f64; }
1485 trait Circle : Shape { fn radius(&self) -> f64; }
1488 the syntax `Circle : Shape` means that types that implement `Circle` must also
1489 have an implementation for `Shape`. Multiple supertraits are separated by `+`,
1490 `trait Circle : Shape + PartialEq { }`. In an implementation of `Circle` for a
1491 given type `T`, methods can refer to `Shape` methods, since the typechecker
1492 checks that any type with an implementation of `Circle` also has an
1493 implementation of `Shape`.
1495 In type-parameterized functions, methods of the supertrait may be called on
1496 values of subtrait-bound type parameters. Referring to the previous example of
1497 `trait Circle : Shape`:
1500 # trait Shape { fn area(&self) -> f64; }
1501 # trait Circle : Shape { fn radius(&self) -> f64; }
1502 fn radius_times_area<T: Circle>(c: T) -> f64 {
1503 // `c` is both a Circle and a Shape
1504 c.radius() * c.area()
1508 Likewise, supertrait methods may also be called on trait objects.
1511 # trait Shape { fn area(&self) -> f64; }
1512 # trait Circle : Shape { fn radius(&self) -> f64; }
1513 # impl Shape for i32 { fn area(&self) -> f64 { 0.0 } }
1514 # impl Circle for i32 { fn radius(&self) -> f64 { 0.0 } }
1515 # let mycircle = 0i32;
1516 let mycircle = Box::new(mycircle) as Box<Circle>;
1517 let nonsense = mycircle.radius() * mycircle.area();
1522 An _implementation_ is an item that implements a [trait](#traits) for a
1525 Implementations are defined with the keyword `impl`.
1528 # #[derive(Copy, Clone)]
1529 # struct Point {x: f64, y: f64};
1530 # type Surface = i32;
1531 # struct BoundingBox {x: f64, y: f64, width: f64, height: f64};
1532 # trait Shape { fn draw(&self, Surface); fn bounding_box(&self) -> BoundingBox; }
1533 # fn do_draw_circle(s: Surface, c: Circle) { }
1539 impl Copy for Circle {}
1541 impl Clone for Circle {
1542 fn clone(&self) -> Circle { *self }
1545 impl Shape for Circle {
1546 fn draw(&self, s: Surface) { do_draw_circle(s, *self); }
1547 fn bounding_box(&self) -> BoundingBox {
1548 let r = self.radius;
1549 BoundingBox{x: self.center.x - r, y: self.center.y - r,
1550 width: 2.0 * r, height: 2.0 * r}
1555 It is possible to define an implementation without referring to a trait. The
1556 methods in such an implementation can only be used as direct calls on the
1557 values of the type that the implementation targets. In such an implementation,
1558 the trait type and `for` after `impl` are omitted. Such implementations are
1559 limited to nominal types (enums, structs), and the implementation must appear
1560 in the same crate as the `self` type:
1563 struct Point {x: i32, y: i32}
1567 println!("Point is at ({}, {})", self.x, self.y);
1571 let my_point = Point {x: 10, y:11};
1575 When a trait _is_ specified in an `impl`, all methods declared as part of the
1576 trait must be implemented, with matching types and type parameter counts.
1578 An implementation can take type parameters, which can be different from the
1579 type parameters taken by the trait it implements. Implementation parameters
1580 are written after the `impl` keyword.
1583 # trait Seq<T> { fn dummy(&self, _: T) { } }
1584 impl<T> Seq<T> for Vec<T> {
1587 impl Seq<bool> for u32 {
1588 /* Treat the integer as a sequence of bits */
1594 External blocks form the basis for Rust's foreign function interface.
1595 Declarations in an external block describe symbols in external, non-Rust
1598 Functions within external blocks are declared in the same way as other Rust
1599 functions, with the exception that they may not have a body and are instead
1600 terminated by a semicolon.
1602 Functions within external blocks may be called by Rust code, just like
1603 functions defined in Rust. The Rust compiler automatically translates between
1604 the Rust ABI and the foreign ABI.
1606 A number of [attributes](#attributes) control the behavior of external blocks.
1608 By default external blocks assume that the library they are calling uses the
1609 standard C "cdecl" ABI. Other ABIs may be specified using an `abi` string, as
1613 // Interface to the Windows API
1614 extern "stdcall" { }
1617 The `link` attribute allows the name of the library to be specified. When
1618 specified the compiler will attempt to link against the native library of the
1622 #[link(name = "crypto")]
1626 The type of a function declared in an extern block is `extern "abi" fn(A1, ...,
1627 An) -> R`, where `A1...An` are the declared types of its arguments and `R` is
1628 the declared return type.
1630 ## Visibility and Privacy
1632 These two terms are often used interchangeably, and what they are attempting to
1633 convey is the answer to the question "Can this item be used at this location?"
1635 Rust's name resolution operates on a global hierarchy of namespaces. Each level
1636 in the hierarchy can be thought of as some item. The items are one of those
1637 mentioned above, but also include external crates. Declaring or defining a new
1638 module can be thought of as inserting a new tree into the hierarchy at the
1639 location of the definition.
1641 To control whether interfaces can be used across modules, Rust checks each use
1642 of an item to see whether it should be allowed or not. This is where privacy
1643 warnings are generated, or otherwise "you used a private item of another module
1644 and weren't allowed to."
1646 By default, everything in Rust is *private*, with one exception. Enum variants
1647 in a `pub` enum are also public by default. When an item is declared as `pub`,
1648 it can be thought of as being accessible to the outside world. For example:
1652 // Declare a private struct
1655 // Declare a public struct with a private field
1660 // Declare a public enum with two public variants
1662 PubliclyAccessibleState,
1663 PubliclyAccessibleState2,
1667 With the notion of an item being either public or private, Rust allows item
1668 accesses in two cases:
1670 1. If an item is public, then it can be used externally through any of its
1672 2. If an item is private, it may be accessed by the current module and its
1675 These two cases are surprisingly powerful for creating module hierarchies
1676 exposing public APIs while hiding internal implementation details. To help
1677 explain, here's a few use cases and what they would entail:
1679 * A library developer needs to expose functionality to crates which link
1680 against their library. As a consequence of the first case, this means that
1681 anything which is usable externally must be `pub` from the root down to the
1682 destination item. Any private item in the chain will disallow external
1685 * A crate needs a global available "helper module" to itself, but it doesn't
1686 want to expose the helper module as a public API. To accomplish this, the
1687 root of the crate's hierarchy would have a private module which then
1688 internally has a "public api". Because the entire crate is a descendant of
1689 the root, then the entire local crate can access this private module through
1692 * When writing unit tests for a module, it's often a common idiom to have an
1693 immediate child of the module to-be-tested named `mod test`. This module
1694 could access any items of the parent module through the second case, meaning
1695 that internal implementation details could also be seamlessly tested from the
1698 In the second case, it mentions that a private item "can be accessed" by the
1699 current module and its descendants, but the exact meaning of accessing an item
1700 depends on what the item is. Accessing a module, for example, would mean
1701 looking inside of it (to import more items). On the other hand, accessing a
1702 function would mean that it is invoked. Additionally, path expressions and
1703 import statements are considered to access an item in the sense that the
1704 import/expression is only valid if the destination is in the current visibility
1707 Here's an example of a program which exemplifies the three cases outlined
1711 // This module is private, meaning that no external crate can access this
1712 // module. Because it is private at the root of this current crate, however, any
1713 // module in the crate may access any publicly visible item in this module.
1714 mod crate_helper_module {
1716 // This function can be used by anything in the current crate
1717 pub fn crate_helper() {}
1719 // This function *cannot* be used by anything else in the crate. It is not
1720 // publicly visible outside of the `crate_helper_module`, so only this
1721 // current module and its descendants may access it.
1722 fn implementation_detail() {}
1725 // This function is "public to the root" meaning that it's available to external
1726 // crates linking against this one.
1727 pub fn public_api() {}
1729 // Similarly to 'public_api', this module is public so external crates may look
1732 use crate_helper_module;
1734 pub fn my_method() {
1735 // Any item in the local crate may invoke the helper module's public
1736 // interface through a combination of the two rules above.
1737 crate_helper_module::crate_helper();
1740 // This function is hidden to any module which is not a descendant of
1742 fn my_implementation() {}
1748 fn test_my_implementation() {
1749 // Because this module is a descendant of `submodule`, it's allowed
1750 // to access private items inside of `submodule` without a privacy
1752 super::my_implementation();
1760 For a rust program to pass the privacy checking pass, all paths must be valid
1761 accesses given the two rules above. This includes all use statements,
1762 expressions, types, etc.
1764 ### Re-exporting and Visibility
1766 Rust allows publicly re-exporting items through a `pub use` directive. Because
1767 this is a public directive, this allows the item to be used in the current
1768 module through the rules above. It essentially allows public access into the
1769 re-exported item. For example, this program is valid:
1772 pub use self::implementation::api;
1774 mod implementation {
1783 This means that any external crate referencing `implementation::api::f` would
1784 receive a privacy violation, while the path `api::f` would be allowed.
1786 When re-exporting a private item, it can be thought of as allowing the "privacy
1787 chain" being short-circuited through the reexport instead of passing through
1788 the namespace hierarchy as it normally would.
1792 Any item declaration may have an _attribute_ applied to it. Attributes in Rust
1793 are modeled on Attributes in ECMA-335, with the syntax coming from ECMA-334
1794 (C#). An attribute is a general, free-form metadatum that is interpreted
1795 according to name, convention, and language and compiler version. Attributes
1796 may appear as any of:
1798 * A single identifier, the attribute name
1799 * An identifier followed by the equals sign '=' and a literal, providing a
1801 * An identifier followed by a parenthesized list of sub-attribute arguments
1803 Attributes with a bang ("!") after the hash ("#") apply to the item that the
1804 attribute is declared within. Attributes that do not have a bang after the hash
1805 apply to the item that follows the attribute.
1807 An example of attributes:
1810 // General metadata applied to the enclosing module or crate.
1811 #![crate_type = "lib"]
1813 // A function marked as a unit test
1819 // A conditionally-compiled module
1820 #[cfg(target_os="linux")]
1825 // A lint attribute used to suppress a warning/error
1826 #[allow(non_camel_case_types)]
1830 > **Note:** At some point in the future, the compiler will distinguish between
1831 > language-reserved and user-available attributes. Until then, there is
1832 > effectively no difference between an attribute handled by a loadable syntax
1833 > extension and the compiler.
1835 ### Crate-only attributes
1837 - `crate_name` - specify the crate's crate name.
1838 - `crate_type` - see [linkage](#linkage).
1839 - `feature` - see [compiler features](#compiler-features).
1840 - `no_builtins` - disable optimizing certain code patterns to invocations of
1841 library functions that are assumed to exist
1842 - `no_main` - disable emitting the `main` symbol. Useful when some other
1843 object being linked to defines `main`.
1844 - `no_start` - disable linking to the `native` crate, which specifies the
1845 "start" language item.
1846 - `no_std` - disable linking to the `std` crate.
1847 - `plugin` — load a list of named crates as compiler plugins, e.g.
1848 `#![plugin(foo, bar)]`. Optional arguments for each plugin,
1849 i.e. `#![plugin(foo(... args ...))]`, are provided to the plugin's
1850 registrar function. The `plugin` feature gate is required to use
1853 ### Module-only attributes
1855 - `no_implicit_prelude` - disable injecting `use std::prelude::*` in this
1857 - `path` - specifies the file to load the module from. `#[path="foo.rs"] mod
1858 bar;` is equivalent to `mod bar { /* contents of foo.rs */ }`. The path is
1859 taken relative to the directory that the current module is in.
1861 ### Function-only attributes
1863 - `main` - indicates that this function should be passed to the entry point,
1864 rather than the function in the crate root named `main`.
1865 - `plugin_registrar` - mark this function as the registration point for
1866 [compiler plugins][plugin], such as loadable syntax extensions.
1867 - `start` - indicates that this function should be used as the entry point,
1868 overriding the "start" language item. See the "start" [language
1869 item](#language-items) for more details.
1870 - `test` - indicates that this function is a test function, to only be compiled
1871 in case of `--test`.
1872 - `should_panic` - indicates that this test function should panic, inverting the success condition.
1873 - `cold` - The function is unlikely to be executed, so optimize it (and calls
1876 ### Static-only attributes
1878 - `thread_local` - on a `static mut`, this signals that the value of this
1879 static may change depending on the current thread. The exact consequences of
1880 this are implementation-defined.
1884 On an `extern` block, the following attributes are interpreted:
1886 - `link_args` - specify arguments to the linker, rather than just the library
1887 name and type. This is feature gated and the exact behavior is
1888 implementation-defined (due to variety of linker invocation syntax).
1889 - `link` - indicate that a native library should be linked to for the
1890 declarations in this block to be linked correctly. `link` supports an optional `kind`
1891 key with three possible values: `dylib`, `static`, and `framework`. See [external blocks](#external-blocks) for more about external blocks. Two
1892 examples: `#[link(name = "readline")]` and
1893 `#[link(name = "CoreFoundation", kind = "framework")]`.
1895 On declarations inside an `extern` block, the following attributes are
1898 - `link_name` - the name of the symbol that this function or static should be
1900 - `linkage` - on a static, this specifies the [linkage
1901 type](http://llvm.org/docs/LangRef.html#linkage-types).
1905 - `repr` - on C-like enums, this sets the underlying type used for
1906 representation. Takes one argument, which is the primitive
1907 type this enum should be represented for, or `C`, which specifies that it
1908 should be the default `enum` size of the C ABI for that platform. Note that
1909 enum representation in C is undefined, and this may be incorrect when the C
1910 code is compiled with certain flags.
1914 - `repr` - specifies the representation to use for this struct. Takes a list
1915 of options. The currently accepted ones are `C` and `packed`, which may be
1916 combined. `C` will use a C ABI compatible struct layout, and `packed` will
1917 remove any padding between fields (note that this is very fragile and may
1918 break platforms which require aligned access).
1920 ### Macro-related attributes
1922 - `macro_use` on a `mod` — macros defined in this module will be visible in the
1923 module's parent, after this module has been included.
1925 - `macro_use` on an `extern crate` — load macros from this crate. An optional
1926 list of names `#[macro_use(foo, bar)]` restricts the import to just those
1927 macros named. The `extern crate` must appear at the crate root, not inside
1928 `mod`, which ensures proper function of the [`$crate` macro
1929 variable](book/macros.html#the-variable-$crate).
1931 - `macro_reexport` on an `extern crate` — re-export the named macros.
1933 - `macro_export` - export a macro for cross-crate usage.
1935 - `no_link` on an `extern crate` — even if we load this crate for macros, don't
1936 link it into the output.
1938 See the [macros section of the
1939 book](book/macros.html#scoping-and-macro-import/export) for more information on
1943 ### Miscellaneous attributes
1945 - `export_name` - on statics and functions, this determines the name of the
1947 - `link_section` - on statics and functions, this specifies the section of the
1948 object file that this item's contents will be placed into.
1949 - `no_mangle` - on any item, do not apply the standard name mangling. Set the
1950 symbol for this item to its identifier.
1951 - `packed` - on structs or enums, eliminate any padding that would be used to
1953 - `simd` - on certain tuple structs, derive the arithmetic operators, which
1954 lower to the target's SIMD instructions, if any; the `simd` feature gate
1955 is necessary to use this attribute.
1956 - `unsafe_no_drop_flag` - on structs, remove the flag that prevents
1957 destructors from being run twice. Destructors might be run multiple times on
1958 the same object with this attribute. To use this, the `unsafe_no_drop_flag` feature
1959 gate must be enabled.
1960 - `doc` - Doc comments such as `/// foo` are equivalent to `#[doc = "foo"]`.
1961 - `rustc_on_unimplemented` - Write a custom note to be shown along with the error
1962 when the trait is found to be unimplemented on a type.
1963 You may use format arguments like `{T}`, `{A}` to correspond to the
1964 types at the point of use corresponding to the type parameters of the
1965 trait of the same name. `{Self}` will be replaced with the type that is supposed
1966 to implement the trait but doesn't. To use this, the `on_unimplemented` feature gate
1969 ### Conditional compilation
1971 Sometimes one wants to have different compiler outputs from the same code,
1972 depending on build target, such as targeted operating system, or to enable
1975 There are two kinds of configuration options, one that is either defined or not
1976 (`#[cfg(foo)]`), and the other that contains a string that can be checked
1977 against (`#[cfg(bar = "baz")]`). Currently, only compiler-defined configuration
1978 options can have the latter form.
1981 // The function is only included in the build when compiling for OSX
1982 #[cfg(target_os = "macos")]
1987 // This function is only included when either foo or bar is defined
1988 #[cfg(any(foo, bar))]
1989 fn needs_foo_or_bar() {
1993 // This function is only included when compiling for a unixish OS with a 32-bit
1995 #[cfg(all(unix, target_pointer_width = "32"))]
1996 fn on_32bit_unix() {
2000 // This function is only included when foo is not defined
2002 fn needs_not_foo() {
2007 This illustrates some conditional compilation can be achieved using the
2008 `#[cfg(...)]` attribute. `any`, `all` and `not` can be used to assemble
2009 arbitrarily complex configurations through nesting.
2011 The following configurations must be defined by the implementation:
2013 * `debug_assertions`. Enabled by default when compiling without optimizations.
2014 This can be used to enable extra debugging code in development but not in
2015 production. For example, it controls the behavior of the standard library's
2016 `debug_assert!` macro.
2017 * `target_arch = "..."`. Target CPU architecture, such as `"x86"`, `"x86_64"`
2018 `"mips"`, `"powerpc"`, `"arm"`, or `"aarch64"`.
2019 * `target_endian = "..."`. Endianness of the target CPU, either `"little"` or
2021 * `target_family = "..."`. Operating system family of the target, e. g.
2022 `"unix"` or `"windows"`. The value of this configuration option is defined
2023 as a configuration itself, like `unix` or `windows`.
2024 * `target_os = "..."`. Operating system of the target, examples include
2025 `"windows"`, `"macos"`, `"ios"`, `"linux"`, `"android"`, `"freebsd"`, `"dragonfly"`,
2026 `"bitrig"` , `"openbsd"` or `"netbsd"`.
2027 * `target_pointer_width = "..."`. Target pointer width in bits. This is set
2028 to `"32"` for targets with 32-bit pointers, and likewise set to `"64"` for
2030 * `unix`. See `target_family`.
2031 * `windows`. See `target_family`.
2033 You can also set another attribute based on a `cfg` variable with `cfg_attr`:
2039 Will be the same as `#[b]` if `a` is set by `cfg`, and nothing otherwise.
2041 ### Lint check attributes
2043 A lint check names a potentially undesirable coding pattern, such as
2044 unreachable code or omitted documentation, for the static entity to which the
2047 For any lint check `C`:
2049 * `allow(C)` overrides the check for `C` so that violations will go
2051 * `deny(C)` signals an error after encountering a violation of `C`,
2052 * `forbid(C)` is the same as `deny(C)`, but also forbids changing the lint
2054 * `warn(C)` warns about violations of `C` but continues compilation.
2056 The lint checks supported by the compiler can be found via `rustc -W help`,
2057 along with their default settings. [Compiler
2058 plugins](book/compiler-plugins.html#lint-plugins) can provide additional lint checks.
2062 // Missing documentation is ignored here
2063 #[allow(missing_docs)]
2064 pub fn undocumented_one() -> i32 { 1 }
2066 // Missing documentation signals a warning here
2067 #[warn(missing_docs)]
2068 pub fn undocumented_too() -> i32 { 2 }
2070 // Missing documentation signals an error here
2071 #[deny(missing_docs)]
2072 pub fn undocumented_end() -> i32 { 3 }
2076 This example shows how one can use `allow` and `warn` to toggle a particular
2080 #[warn(missing_docs)]
2082 #[allow(missing_docs)]
2084 // Missing documentation is ignored here
2085 pub fn undocumented_one() -> i32 { 1 }
2087 // Missing documentation signals a warning here,
2088 // despite the allow above.
2089 #[warn(missing_docs)]
2090 pub fn undocumented_two() -> i32 { 2 }
2093 // Missing documentation signals a warning here
2094 pub fn undocumented_too() -> i32 { 3 }
2098 This example shows how one can use `forbid` to disallow uses of `allow` for
2102 #[forbid(missing_docs)]
2104 // Attempting to toggle warning signals an error here
2105 #[allow(missing_docs)]
2107 pub fn undocumented_too() -> i32 { 2 }
2113 Some primitive Rust operations are defined in Rust code, rather than being
2114 implemented directly in C or assembly language. The definitions of these
2115 operations have to be easy for the compiler to find. The `lang` attribute
2116 makes it possible to declare these operations. For example, the `str` module
2117 in the Rust standard library defines the string equality function:
2121 pub fn eq_slice(a: &str, b: &str) -> bool {
2126 The name `str_eq` has a special meaning to the Rust compiler, and the presence
2127 of this definition means that it will use this definition when generating calls
2128 to the string equality function.
2130 The set of language items is currently considered unstable. A complete
2131 list of the built-in language items will be added in the future.
2133 ### Inline attributes
2135 The inline attribute suggests that the compiler should place a copy of
2136 the function or static in the caller, rather than generating code to
2137 call the function or access the static where it is defined.
2139 The compiler automatically inlines functions based on internal heuristics.
2140 Incorrectly inlining functions can actually make the program slower, so it
2141 should be used with care.
2143 `#[inline]` and `#[inline(always)]` always cause the function to be serialized
2144 into the crate metadata to allow cross-crate inlining.
2146 There are three different types of inline attributes:
2148 * `#[inline]` hints the compiler to perform an inline expansion.
2149 * `#[inline(always)]` asks the compiler to always perform an inline expansion.
2150 * `#[inline(never)]` asks the compiler to never perform an inline expansion.
2154 The `derive` attribute allows certain traits to be automatically implemented
2155 for data structures. For example, the following will create an `impl` for the
2156 `PartialEq` and `Clone` traits for `Foo`, the type parameter `T` will be given
2157 the `PartialEq` or `Clone` constraints for the appropriate `impl`:
2160 #[derive(PartialEq, Clone)]
2167 The generated `impl` for `PartialEq` is equivalent to
2170 # struct Foo<T> { a: i32, b: T }
2171 impl<T: PartialEq> PartialEq for Foo<T> {
2172 fn eq(&self, other: &Foo<T>) -> bool {
2173 self.a == other.a && self.b == other.b
2176 fn ne(&self, other: &Foo<T>) -> bool {
2177 self.a != other.a || self.b != other.b
2182 ### Compiler Features
2184 Certain aspects of Rust may be implemented in the compiler, but they're not
2185 necessarily ready for every-day use. These features are often of "prototype
2186 quality" or "almost production ready", but may not be stable enough to be
2187 considered a full-fledged language feature.
2189 For this reason, Rust recognizes a special crate-level attribute of the form:
2192 #![feature(feature1, feature2, feature3)]
2195 This directive informs the compiler that the feature list: `feature1`,
2196 `feature2`, and `feature3` should all be enabled. This is only recognized at a
2197 crate-level, not at a module-level. Without this directive, all features are
2198 considered off, and using the features will result in a compiler error.
2200 The currently implemented features of the reference compiler are:
2202 * `advanced_slice_patterns` - See the [match expressions](#match-expressions)
2203 section for discussion; the exact semantics of
2204 slice patterns are subject to change, so some types
2207 * `slice_patterns` - OK, actually, slice patterns are just scary and
2208 completely unstable.
2210 * `asm` - The `asm!` macro provides a means for inline assembly. This is often
2211 useful, but the exact syntax for this feature along with its
2212 semantics are likely to change, so this macro usage must be opted
2215 * `associated_consts` - Allows constants to be defined in `impl` and `trait`
2216 blocks, so that they can be associated with a type or
2217 trait in a similar manner to methods and associated
2220 * `box_patterns` - Allows `box` patterns, the exact semantics of which
2221 is subject to change.
2223 * `box_syntax` - Allows use of `box` expressions, the exact semantics of which
2224 is subject to change.
2226 * `concat_idents` - Allows use of the `concat_idents` macro, which is in many
2227 ways insufficient for concatenating identifiers, and may be
2228 removed entirely for something more wholesome.
2230 * `custom_attribute` - Allows the usage of attributes unknown to the compiler
2231 so that new attributes can be added in a backwards compatible
2234 * `custom_derive` - Allows the use of `#[derive(Foo,Bar)]` as sugar for
2235 `#[derive_Foo] #[derive_Bar]`, which can be user-defined syntax
2238 * `intrinsics` - Allows use of the "rust-intrinsics" ABI. Compiler intrinsics
2239 are inherently unstable and no promise about them is made.
2241 * `lang_items` - Allows use of the `#[lang]` attribute. Like `intrinsics`,
2242 lang items are inherently unstable and no promise about them
2245 * `link_args` - This attribute is used to specify custom flags to the linker,
2246 but usage is strongly discouraged. The compiler's usage of the
2247 system linker is not guaranteed to continue in the future, and
2248 if the system linker is not used then specifying custom flags
2249 doesn't have much meaning.
2251 * `link_llvm_intrinsics` – Allows linking to LLVM intrinsics via
2252 `#[link_name="llvm.*"]`.
2254 * `linkage` - Allows use of the `linkage` attribute, which is not portable.
2256 * `log_syntax` - Allows use of the `log_syntax` macro attribute, which is a
2257 nasty hack that will certainly be removed.
2259 * `main` - Allows use of the `#[main]` attribute, which changes the entry point
2260 into a Rust program. This capability is subject to change.
2262 * `macro_reexport` - Allows macros to be re-exported from one crate after being imported
2263 from another. This feature was originally designed with the sole
2264 use case of the Rust standard library in mind, and is subject to
2267 * `non_ascii_idents` - The compiler supports the use of non-ascii identifiers,
2268 but the implementation is a little rough around the
2269 edges, so this can be seen as an experimental feature
2270 for now until the specification of identifiers is fully
2273 * `no_std` - Allows the `#![no_std]` crate attribute, which disables the implicit
2274 `extern crate std`. This typically requires use of the unstable APIs
2275 behind the libstd "facade", such as libcore and libcollections. It
2276 may also cause problems when using syntax extensions, including
2279 * `on_unimplemented` - Allows the `#[rustc_on_unimplemented]` attribute, which allows
2280 trait definitions to add specialized notes to error messages
2281 when an implementation was expected but not found.
2283 * `optin_builtin_traits` - Allows the definition of default and negative trait
2284 implementations. Experimental.
2286 * `plugin` - Usage of [compiler plugins][plugin] for custom lints or syntax extensions.
2287 These depend on compiler internals and are subject to change.
2289 * `plugin_registrar` - Indicates that a crate provides [compiler plugins][plugin].
2291 * `quote` - Allows use of the `quote_*!` family of macros, which are
2292 implemented very poorly and will likely change significantly
2293 with a proper implementation.
2295 * `rustc_attrs` - Gates internal `#[rustc_*]` attributes which may be
2296 for internal use only or have meaning added to them in the future.
2298 * `rustc_diagnostic_macros`- A mysterious feature, used in the implementation
2299 of rustc, not meant for mortals.
2301 * `simd` - Allows use of the `#[simd]` attribute, which is overly simple and
2302 not the SIMD interface we want to expose in the long term.
2304 * `simd_ffi` - Allows use of SIMD vectors in signatures for foreign functions.
2305 The SIMD interface is subject to change.
2307 * `staged_api` - Allows usage of stability markers and `#![staged_api]` in a
2308 crate. Stability markers are also attributes: `#[stable]`,
2309 `#[unstable]`, and `#[deprecated]` are the three levels.
2311 * `start` - Allows use of the `#[start]` attribute, which changes the entry point
2312 into a Rust program. This capability, especially the signature for the
2313 annotated function, is subject to change.
2315 * `struct_inherit` - Allows using struct inheritance, which is barely
2316 implemented and will probably be removed. Don't use this.
2318 * `struct_variant` - Structural enum variants (those with named fields). It is
2319 currently unknown whether this style of enum variant is as
2320 fully supported as the tuple-forms, and it's not certain
2321 that this style of variant should remain in the language.
2322 For now this style of variant is hidden behind a feature
2325 * `thread_local` - The usage of the `#[thread_local]` attribute is experimental
2326 and should be seen as unstable. This attribute is used to
2327 declare a `static` as being unique per-thread leveraging
2328 LLVM's implementation which works in concert with the kernel
2329 loader and dynamic linker. This is not necessarily available
2330 on all platforms, and usage of it is discouraged.
2332 * `trace_macros` - Allows use of the `trace_macros` macro, which is a nasty
2333 hack that will certainly be removed.
2335 * `unboxed_closures` - Rust's new closure design, which is currently a work in
2336 progress feature with many known bugs.
2338 * `unsafe_no_drop_flag` - Allows use of the `#[unsafe_no_drop_flag]` attribute,
2339 which removes hidden flag added to a type that
2340 implements the `Drop` trait. The design for the
2341 `Drop` flag is subject to change, and this feature
2342 may be removed in the future.
2344 * `unmarked_api` - Allows use of items within a `#![staged_api]` crate
2345 which have not been marked with a stability marker.
2346 Such items should not be allowed by the compiler to exist,
2347 so if you need this there probably is a compiler bug.
2349 * `visible_private_types` - Allows public APIs to expose otherwise private
2350 types, e.g. as the return type of a public function.
2351 This capability may be removed in the future.
2353 * `allow_internal_unstable` - Allows `macro_rules!` macros to be tagged with the
2354 `#[allow_internal_unstable]` attribute, designed
2355 to allow `std` macros to call
2356 `#[unstable]`/feature-gated functionality
2357 internally without imposing on callers
2358 (i.e. making them behave like function calls in
2359 terms of encapsulation).
2361 If a feature is promoted to a language feature, then all existing programs will
2362 start to receive compilation warnings about `#![feature]` directives which enabled
2363 the new feature (because the directive is no longer necessary). However, if a
2364 feature is decided to be removed from the language, errors will be issued (if
2365 there isn't a parser error first). The directive in this case is no longer
2366 necessary, and it's likely that existing code will break if the feature isn't
2369 If an unknown feature is found in a directive, it results in a compiler error.
2370 An unknown feature is one which has never been recognized by the compiler.
2372 # Statements and expressions
2374 Rust is _primarily_ an expression language. This means that most forms of
2375 value-producing or effect-causing evaluation are directed by the uniform syntax
2376 category of _expressions_. Each kind of expression can typically _nest_ within
2377 each other kind of expression, and rules for evaluation of expressions involve
2378 specifying both the value produced by the expression and the order in which its
2379 sub-expressions are themselves evaluated.
2381 In contrast, statements in Rust serve _mostly_ to contain and explicitly
2382 sequence expression evaluation.
2386 A _statement_ is a component of a block, which is in turn a component of an
2387 outer [expression](#expressions) or [function](#functions).
2389 Rust has two kinds of statement: [declaration
2390 statements](#declaration-statements) and [expression
2391 statements](#expression-statements).
2393 ### Declaration statements
2395 A _declaration statement_ is one that introduces one or more *names* into the
2396 enclosing statement block. The declared names may denote new variables or new
2399 #### Item declarations
2401 An _item declaration statement_ has a syntactic form identical to an
2402 [item](#items) declaration within a module. Declaring an item — a
2403 function, enumeration, structure, type, static, trait, implementation or module
2404 — locally within a statement block is simply a way of restricting its
2405 scope to a narrow region containing all of its uses; it is otherwise identical
2406 in meaning to declaring the item outside the statement block.
2408 > **Note**: there is no implicit capture of the function's dynamic environment when
2409 > declaring a function-local item.
2411 #### Variable declarations
2413 A _variable declaration_ introduces a new set of variable, given by a pattern. The
2414 pattern may be followed by a type annotation, and/or an initializer expression.
2415 When no type annotation is given, the compiler will infer the type, or signal
2416 an error if insufficient type information is available for definite inference.
2417 Any variables introduced by a variable declaration are visible from the point of
2418 declaration until the end of the enclosing block scope.
2420 ### Expression statements
2422 An _expression statement_ is one that evaluates an [expression](#expressions)
2423 and ignores its result. The type of an expression statement `e;` is always
2424 `()`, regardless of the type of `e`. As a rule, an expression statement's
2425 purpose is to trigger the effects of evaluating its expression.
2429 An expression may have two roles: it always produces a *value*, and it may have
2430 *effects* (otherwise known as "side effects"). An expression *evaluates to* a
2431 value, and has effects during *evaluation*. Many expressions contain
2432 sub-expressions (operands). The meaning of each kind of expression dictates
2435 * Whether or not to evaluate the sub-expressions when evaluating the expression
2436 * The order in which to evaluate the sub-expressions
2437 * How to combine the sub-expressions' values to obtain the value of the expression
2439 In this way, the structure of expressions dictates the structure of execution.
2440 Blocks are just another kind of expression, so blocks, statements, expressions,
2441 and blocks again can recursively nest inside each other to an arbitrary depth.
2443 #### Lvalues, rvalues and temporaries
2445 Expressions are divided into two main categories: _lvalues_ and _rvalues_.
2446 Likewise within each expression, sub-expressions may occur in _lvalue context_
2447 or _rvalue context_. The evaluation of an expression depends both on its own
2448 category and the context it occurs within.
2450 An lvalue is an expression that represents a memory location. These expressions
2451 are [paths](#path-expressions) (which refer to local variables, function and
2452 method arguments, or static variables), dereferences (`*expr`), [indexing
2453 expressions](#index-expressions) (`expr[expr]`), and [field
2454 references](#field-expressions) (`expr.f`). All other expressions are rvalues.
2456 The left operand of an [assignment](#assignment-expressions) or
2457 [compound-assignment](#compound-assignment-expressions) expression is
2458 an lvalue context, as is the single operand of a unary
2459 [borrow](#unary-operator-expressions). The discriminant or subject of
2460 a [match expression](#match-expressions) may be an lvalue context, if
2461 ref bindings are made, but is otherwise an rvalue context. All other
2462 expression contexts are rvalue contexts.
2464 When an lvalue is evaluated in an _lvalue context_, it denotes a memory
2465 location; when evaluated in an _rvalue context_, it denotes the value held _in_
2466 that memory location.
2468 ##### Temporary lifetimes
2470 When an rvalue is used in an lvalue context, a temporary un-named
2471 lvalue is created and used instead. The lifetime of temporary values
2472 is typically the innermost enclosing statement; the tail expression of
2473 a block is considered part of the statement that encloses the block.
2475 When a temporary rvalue is being created that is assigned into a `let`
2476 declaration, however, the temporary is created with the lifetime of
2477 the enclosing block instead, as using the enclosing statement (the
2478 `let` declaration) would be a guaranteed error (since a pointer to the
2479 temporary would be stored into a variable, but the temporary would be
2480 freed before the variable could be used). The compiler uses simple
2481 syntactic rules to decide which values are being assigned into a `let`
2482 binding, and therefore deserve a longer temporary lifetime.
2484 Here are some examples:
2486 - `let x = foo(&temp())`. The expression `temp()` is an rvalue. As it
2487 is being borrowed, a temporary is created which will be freed after
2488 the innermost enclosing statement (the `let` declaration, in this case).
2489 - `let x = temp().foo()`. This is the same as the previous example,
2490 except that the value of `temp()` is being borrowed via autoref on a
2491 method-call. Here we are assuming that `foo()` is an `&self` method
2492 defined in some trait, say `Foo`. In other words, the expression
2493 `temp().foo()` is equivalent to `Foo::foo(&temp())`.
2494 - `let x = &temp()`. Here, the same temporary is being assigned into
2495 `x`, rather than being passed as a parameter, and hence the
2496 temporary's lifetime is considered to be the enclosing block.
2497 - `let x = SomeStruct { foo: &temp() }`. As in the previous case, the
2498 temporary is assigned into a struct which is then assigned into a
2499 binding, and hence it is given the lifetime of the enclosing block.
2500 - `let x = [ &temp() ]`. As in the previous case, the
2501 temporary is assigned into an array which is then assigned into a
2502 binding, and hence it is given the lifetime of the enclosing block.
2503 - `let ref x = temp()`. In this case, the temporary is created using a ref binding,
2504 but the result is the same: the lifetime is extended to the enclosing block.
2506 #### Moved and copied types
2508 When a [local variable](#variables) is used as an
2509 [rvalue](#lvalues,-rvalues-and-temporaries) the variable will either be moved
2510 or copied, depending on its type. All values whose type implements `Copy` are
2511 copied, all others are moved.
2513 ### Literal expressions
2515 A _literal expression_ consists of one of the [literal](#literals) forms
2516 described earlier. It directly describes a number, character, string, boolean
2517 value, or the unit value.
2521 "hello"; // string type
2522 '5'; // character type
2526 ### Path expressions
2528 A [path](#paths) used as an expression context denotes either a local variable
2529 or an item. Path expressions are [lvalues](#lvalues,-rvalues-and-temporaries).
2531 ### Tuple expressions
2533 Tuples are written by enclosing zero or more comma-separated expressions in
2534 parentheses. They are used to create [tuple-typed](#tuple-types) values.
2538 ("a", 4usize, true);
2541 You can disambiguate a single-element tuple from a value in parentheses with a
2545 (0,); // single-element tuple
2546 (0); // zero in parentheses
2549 ### Structure expressions
2551 There are several forms of structure expressions. A _structure expression_
2552 consists of the [path](#paths) of a [structure item](#structures), followed by
2553 a brace-enclosed list of one or more comma-separated name-value pairs,
2554 providing the field values of a new instance of the structure. A field name
2555 can be any identifier, and is separated from its value expression by a colon.
2556 The location denoted by a structure field is mutable if and only if the
2557 enclosing structure is mutable.
2559 A _tuple structure expression_ consists of the [path](#paths) of a [structure
2560 item](#structures), followed by a parenthesized list of one or more
2561 comma-separated expressions (in other words, the path of a structure item
2562 followed by a tuple expression). The structure item must be a tuple structure
2565 A _unit-like structure expression_ consists only of the [path](#paths) of a
2566 [structure item](#structures).
2568 The following are examples of structure expressions:
2571 # struct Point { x: f64, y: f64 }
2572 # struct TuplePoint(f64, f64);
2573 # mod game { pub struct User<'a> { pub name: &'a str, pub age: u32, pub score: usize } }
2574 # struct Cookie; fn some_fn<T>(t: T) {}
2575 Point {x: 10.0, y: 20.0};
2576 TuplePoint(10.0, 20.0);
2577 let u = game::User {name: "Joe", age: 35, score: 100_000};
2578 some_fn::<Cookie>(Cookie);
2581 A structure expression forms a new value of the named structure type. Note
2582 that for a given *unit-like* structure type, this will always be the same
2585 A structure expression can terminate with the syntax `..` followed by an
2586 expression to denote a functional update. The expression following `..` (the
2587 base) must have the same structure type as the new structure type being formed.
2588 The entire expression denotes the result of constructing a new structure (with
2589 the same type as the base expression) with the given values for the fields that
2590 were explicitly specified and the values in the base expression for all other
2594 # struct Point3d { x: i32, y: i32, z: i32 }
2595 let base = Point3d {x: 1, y: 2, z: 3};
2596 Point3d {y: 0, z: 10, .. base};
2599 ### Block expressions
2601 A _block expression_ is similar to a module in terms of the declarations that
2602 are possible. Each block conceptually introduces a new namespace scope. Use
2603 items can bring new names into scopes and declared items are in scope for only
2606 A block will execute each statement sequentially, and then execute the
2607 expression (if given). If the block ends in a statement, its value is `()`:
2610 let x: () = { println!("Hello."); };
2613 If it ends in an expression, its value and type are that of the expression:
2616 let x: i32 = { println!("Hello."); 5 };
2621 ### Method-call expressions
2623 A _method call_ consists of an expression followed by a single dot, an
2624 identifier, and a parenthesized expression-list. Method calls are resolved to
2625 methods on specific traits, either statically dispatching to a method if the
2626 exact `self`-type of the left-hand-side is known, or dynamically dispatching if
2627 the left-hand-side expression is an indirect [trait object](#trait-objects).
2629 ### Field expressions
2631 A _field expression_ consists of an expression followed by a single dot and an
2632 identifier, when not immediately followed by a parenthesized expression-list
2633 (the latter is a [method call expression](#method-call-expressions)). A field
2634 expression denotes a field of a [structure](#structure-types).
2639 (Struct {a: 10, b: 20}).a;
2642 A field access is an [lvalue](#lvalues,-rvalues-and-temporaries) referring to
2643 the value of that field. When the type providing the field inherits mutability,
2644 it can be [assigned](#assignment-expressions) to.
2646 Also, if the type of the expression to the left of the dot is a
2647 pointer, it is automatically dereferenced as many times as necessary
2648 to make the field access possible. In cases of ambiguity, we prefer
2649 fewer autoderefs to more.
2651 ### Array expressions
2653 An [array](#array,-and-slice-types) _expression_ is written by enclosing zero
2654 or more comma-separated expressions of uniform type in square brackets.
2656 In the `[expr ';' expr]` form, the expression after the `';'` must be a
2657 constant expression that can be evaluated at compile time, such as a
2658 [literal](#literals) or a [static item](#static-items).
2662 ["a", "b", "c", "d"];
2663 [0; 128]; // array with 128 zeros
2664 [0u8, 0u8, 0u8, 0u8];
2667 ### Index expressions
2669 [Array](#array,-and-slice-types)-typed expressions can be indexed by
2670 writing a square-bracket-enclosed expression (the index) after them. When the
2671 array is mutable, the resulting [lvalue](#lvalues,-rvalues-and-temporaries) can
2674 Indices are zero-based, and may be of any integral type. Vector access is
2675 bounds-checked at compile-time for constant arrays being accessed with a constant index value.
2676 Otherwise a check will be performed at run-time that will put the thread in a _panicked state_ if it fails.
2681 let x = (["a", "b"])[10]; // compiler error: const index-expr is out of bounds
2684 let y = (["a", "b"])[n]; // panics
2686 let arr = ["a", "b"];
2690 Also, if the type of the expression to the left of the brackets is a
2691 pointer, it is automatically dereferenced as many times as necessary
2692 to make the indexing possible. In cases of ambiguity, we prefer fewer
2695 ### Range expressions
2697 The `..` operator will construct an object of one of the `std::ops::Range` variants.
2700 1..2; // std::ops::Range
2701 3..; // std::ops::RangeFrom
2702 ..4; // std::ops::RangeTo
2703 ..; // std::ops::RangeFull
2706 The following expressions are equivalent.
2709 let x = std::ops::Range {start: 0, end: 10};
2715 ### Unary operator expressions
2717 Rust defines the following unary operators. They are all written as prefix operators,
2718 before the expression they apply to.
2721 : Negation. May only be applied to numeric types.
2723 : Dereference. When applied to a [pointer](#pointer-types) it denotes the
2724 pointed-to location. For pointers to mutable locations, the resulting
2725 [lvalue](#lvalues,-rvalues-and-temporaries) can be assigned to.
2726 On non-pointer types, it calls the `deref` method of the `std::ops::Deref`
2727 trait, or the `deref_mut` method of the `std::ops::DerefMut` trait (if
2728 implemented by the type and required for an outer expression that will or
2729 could mutate the dereference), and produces the result of dereferencing the
2730 `&` or `&mut` borrowed pointer returned from the overload method.
2732 : Logical negation. On the boolean type, this flips between `true` and
2733 `false`. On integer types, this inverts the individual bits in the
2734 two's complement representation of the value.
2736 : Borrowing. When applied to an lvalue, these operators produce a
2737 reference (pointer) to the lvalue. The lvalue is also placed into
2738 a borrowed state for the duration of the reference. For a shared
2739 borrow (`&`), this implies that the lvalue may not be mutated, but
2740 it may be read or shared again. For a mutable borrow (`&mut`), the
2741 lvalue may not be accessed in any way until the borrow expires.
2742 If the `&` or `&mut` operators are applied to an rvalue, a
2743 temporary value is created; the lifetime of this temporary value
2744 is defined by [syntactic rules](#temporary-lifetimes).
2746 ### Binary operator expressions
2748 Binary operators expressions are given in terms of [operator
2749 precedence](#operator-precedence).
2751 #### Arithmetic operators
2753 Binary arithmetic expressions are syntactic sugar for calls to built-in traits,
2754 defined in the `std::ops` module of the `std` library. This means that
2755 arithmetic operators can be overridden for user-defined types. The default
2756 meaning of the operators on standard types is given here.
2759 : Addition and array/string concatenation.
2760 Calls the `add` method on the `std::ops::Add` trait.
2763 Calls the `sub` method on the `std::ops::Sub` trait.
2766 Calls the `mul` method on the `std::ops::Mul` trait.
2769 Calls the `div` method on the `std::ops::Div` trait.
2772 Calls the `rem` method on the `std::ops::Rem` trait.
2774 #### Bitwise operators
2776 Like the [arithmetic operators](#arithmetic-operators), bitwise operators are
2777 syntactic sugar for calls to methods of built-in traits. This means that
2778 bitwise operators can be overridden for user-defined types. The default
2779 meaning of the operators on standard types is given here. Bitwise `&`, `|` and
2780 `^` applied to boolean arguments are equivalent to logical `&&`, `||` and `!=`
2781 evaluated in non-lazy fashion.
2785 Calls the `bitand` method of the `std::ops::BitAnd` trait.
2787 : Bitwise inclusive OR.
2788 Calls the `bitor` method of the `std::ops::BitOr` trait.
2790 : Bitwise exclusive OR.
2791 Calls the `bitxor` method of the `std::ops::BitXor` trait.
2794 Calls the `shl` method of the `std::ops::Shl` trait.
2796 : Right shift (arithmetic).
2797 Calls the `shr` method of the `std::ops::Shr` trait.
2799 #### Lazy boolean operators
2801 The operators `||` and `&&` may be applied to operands of boolean type. The
2802 `||` operator denotes logical 'or', and the `&&` operator denotes logical
2803 'and'. They differ from `|` and `&` in that the right-hand operand is only
2804 evaluated when the left-hand operand does not already determine the result of
2805 the expression. That is, `||` only evaluates its right-hand operand when the
2806 left-hand operand evaluates to `false`, and `&&` only when it evaluates to
2809 #### Comparison operators
2811 Comparison operators are, like the [arithmetic
2812 operators](#arithmetic-operators), and [bitwise operators](#bitwise-operators),
2813 syntactic sugar for calls to built-in traits. This means that comparison
2814 operators can be overridden for user-defined types. The default meaning of the
2815 operators on standard types is given here.
2819 Calls the `eq` method on the `std::cmp::PartialEq` trait.
2822 Calls the `ne` method on the `std::cmp::PartialEq` trait.
2825 Calls the `lt` method on the `std::cmp::PartialOrd` trait.
2828 Calls the `gt` method on the `std::cmp::PartialOrd` trait.
2830 : Less than or equal.
2831 Calls the `le` method on the `std::cmp::PartialOrd` trait.
2833 : Greater than or equal.
2834 Calls the `ge` method on the `std::cmp::PartialOrd` trait.
2836 #### Type cast expressions
2838 A type cast expression is denoted with the binary operator `as`.
2840 Executing an `as` expression casts the value on the left-hand side to the type
2841 on the right-hand side.
2843 An example of an `as` expression:
2846 # fn sum(values: &[f64]) -> f64 { 0.0 }
2847 # fn len(values: &[f64]) -> i32 { 0 }
2849 fn average(values: &[f64]) -> f64 {
2850 let sum: f64 = sum(values);
2851 let size: f64 = len(values) as f64;
2856 Some of the conversions which can be done through the `as` operator
2857 can also be done implicitly at various points in the program, such as
2858 argument passing and assignment to a `let` binding with an explicit
2859 type. Implicit conversions are limited to "harmless" conversions that
2860 do not lose information and which have minimal or no risk of
2861 surprising side-effects on the dynamic execution semantics.
2863 #### Assignment expressions
2865 An _assignment expression_ consists of an
2866 [lvalue](#lvalues,-rvalues-and-temporaries) expression followed by an equals
2867 sign (`=`) and an [rvalue](#lvalues,-rvalues-and-temporaries) expression.
2869 Evaluating an assignment expression [either copies or
2870 moves](#moved-and-copied-types) its right-hand operand to its left-hand
2880 #### Compound assignment expressions
2882 The `+`, `-`, `*`, `/`, `%`, `&`, `|`, `^`, `<<`, and `>>` operators may be
2883 composed with the `=` operator. The expression `lval OP= val` is equivalent to
2884 `lval = lval OP val`. For example, `x = x + 1` may be written as `x += 1`.
2886 Any such expression always has the [`unit`](#tuple-types) type.
2888 #### Operator precedence
2890 The precedence of Rust binary operators is ordered as follows, going from
2893 ```{.text .precedence}
2907 Operators at the same precedence level are evaluated left-to-right. [Unary
2908 operators](#unary-operator-expressions) have the same precedence level and are
2909 stronger than any of the binary operators.
2911 ### Grouped expressions
2913 An expression enclosed in parentheses evaluates to the result of the enclosed
2914 expression. Parentheses can be used to explicitly specify evaluation order
2915 within an expression.
2917 An example of a parenthesized expression:
2920 let x: i32 = (2 + 3) * 4;
2924 ### Call expressions
2926 A _call expression_ invokes a function, providing zero or more input variables
2927 and an optional location to move the function's output into. If the function
2928 eventually returns, then the expression completes.
2930 Some examples of call expressions:
2933 # fn add(x: i32, y: i32) -> i32 { 0 }
2935 let x: i32 = add(1i32, 2i32);
2936 let pi: Result<f32, _> = "3.14".parse();
2939 ### Lambda expressions
2941 A _lambda expression_ (sometimes called an "anonymous function expression")
2942 defines a function and denotes it as a value, in a single expression. A lambda
2943 expression is a pipe-symbol-delimited (`|`) list of identifiers followed by an
2946 A lambda expression denotes a function that maps a list of parameters
2947 (`ident_list`) onto the expression that follows the `ident_list`. The
2948 identifiers in the `ident_list` are the parameters to the function. These
2949 parameters' types need not be specified, as the compiler infers them from
2952 Lambda expressions are most useful when passing functions as arguments to other
2953 functions, as an abbreviation for defining and capturing a separate function.
2955 Significantly, lambda expressions _capture their environment_, which regular
2956 [function definitions](#functions) do not. The exact type of capture depends
2957 on the [function type](#function-types) inferred for the lambda expression. In
2958 the simplest and least-expensive form (analogous to a ```|| { }``` expression),
2959 the lambda expression captures its environment by reference, effectively
2960 borrowing pointers to all outer variables mentioned inside the function.
2961 Alternately, the compiler may infer that a lambda expression should copy or
2962 move values (depending on their type) from the environment into the lambda
2963 expression's captured environment.
2965 In this example, we define a function `ten_times` that takes a higher-order
2966 function argument, and call it with a lambda expression as an argument:
2969 fn ten_times<F>(f: F) where F: Fn(i32) {
2977 ten_times(|j| println!("hello, {}", j));
2982 A `loop` expression denotes an infinite loop.
2984 A `loop` expression may optionally have a _label_. The label is written as
2985 a lifetime preceding the loop expression, as in `'foo: loop{ }`. If a
2986 label is present, then labeled `break` and `continue` expressions nested
2987 within this loop may exit out of this loop or return control to its head.
2988 See [Break expressions](#break-expressions) and [Continue
2989 expressions](#continue-expressions).
2991 ### Break expressions
2993 A `break` expression has an optional _label_. If the label is absent, then
2994 executing a `break` expression immediately terminates the innermost loop
2995 enclosing it. It is only permitted in the body of a loop. If the label is
2996 present, then `break 'foo` terminates the loop with label `'foo`, which need not
2997 be the innermost label enclosing the `break` expression, but must enclose it.
2999 ### Continue expressions
3001 A `continue` expression has an optional _label_. If the label is absent, then
3002 executing a `continue` expression immediately terminates the current iteration
3003 of the innermost loop enclosing it, returning control to the loop *head*. In
3004 the case of a `while` loop, the head is the conditional expression controlling
3005 the loop. In the case of a `for` loop, the head is the call-expression
3006 controlling the loop. If the label is present, then `continue 'foo` returns
3007 control to the head of the loop with label `'foo`, which need not be the
3008 innermost label enclosing the `break` expression, but must enclose it.
3010 A `continue` expression is only permitted in the body of a loop.
3014 A `while` loop begins by evaluating the boolean loop conditional expression.
3015 If the loop conditional expression evaluates to `true`, the loop body block
3016 executes and control returns to the loop conditional expression. If the loop
3017 conditional expression evaluates to `false`, the `while` expression completes.
3030 Like `loop` expressions, `while` loops can be controlled with `break` or
3031 `continue`, and may optionally have a _label_. See [infinite
3032 loops](#infinite-loops), [break expressions](#break-expressions), and
3033 [continue expressions](#continue-expressions) for more information.
3037 A `for` expression is a syntactic construct for looping over elements provided
3038 by an implementation of `std::iter::IntoIterator`.
3040 An example of a for loop over the contents of an array:
3044 # fn bar(f: &Foo) { }
3049 let v: &[Foo] = &[a, b, c];
3056 An example of a for loop over a series of integers:
3059 # fn bar(b:usize) { }
3065 Like `loop` expressions, `for` loops can be controlled with `break` or
3066 `continue`, and may optionally have a _label_. See [infinite
3067 loops](#infinite-loops), [break expressions](#break-expressions), and
3068 [continue expressions](#continue-expressions) for more information.
3072 An `if` expression is a conditional branch in program control. The form of an
3073 `if` expression is a condition expression, followed by a consequent block, any
3074 number of `else if` conditions and blocks, and an optional trailing `else`
3075 block. The condition expressions must have type `bool`. If a condition
3076 expression evaluates to `true`, the consequent block is executed and any
3077 subsequent `else if` or `else` block is skipped. If a condition expression
3078 evaluates to `false`, the consequent block is skipped and any subsequent `else
3079 if` condition is evaluated. If all `if` and `else if` conditions evaluate to
3080 `false` then any `else` block is executed.
3082 ### Match expressions
3084 A `match` expression branches on a *pattern*. The exact form of matching that
3085 occurs depends on the pattern. Patterns consist of some combination of
3086 literals, destructured arrays or enum constructors, structures and tuples,
3087 variable binding specifications, wildcards (`..`), and placeholders (`_`). A
3088 `match` expression has a *head expression*, which is the value to compare to
3089 the patterns. The type of the patterns must equal the type of the head
3092 In a pattern whose head expression has an `enum` type, a placeholder (`_`)
3093 stands for a *single* data field, whereas a wildcard `..` stands for *all* the
3094 fields of a particular variant.
3096 A `match` behaves differently depending on whether or not the head expression
3097 is an [lvalue or an rvalue](#lvalues,-rvalues-and-temporaries). If the head
3098 expression is an rvalue, it is first evaluated into a temporary location, and
3099 the resulting value is sequentially compared to the patterns in the arms until
3100 a match is found. The first arm with a matching pattern is chosen as the branch
3101 target of the `match`, any variables bound by the pattern are assigned to local
3102 variables in the arm's block, and control enters the block.
3104 When the head expression is an lvalue, the match does not allocate a temporary
3105 location (however, a by-value binding may copy or move from the lvalue). When
3106 possible, it is preferable to match on lvalues, as the lifetime of these
3107 matches inherits the lifetime of the lvalue, rather than being restricted to
3108 the inside of the match.
3110 An example of a `match` expression:
3116 1 => println!("one"),
3117 2 => println!("two"),
3118 3 => println!("three"),
3119 4 => println!("four"),
3120 5 => println!("five"),
3121 _ => println!("something else"),
3125 Patterns that bind variables default to binding to a copy or move of the
3126 matched value (depending on the matched value's type). This can be changed to
3127 bind to a reference by using the `ref` keyword, or to a mutable reference using
3130 Subpatterns can also be bound to variables by the use of the syntax `variable @
3131 subpattern`. For example:
3137 e @ 1 ... 5 => println!("got a range element {}", e),
3138 _ => println!("anything"),
3142 Patterns can also dereference pointers by using the `&`, `&mut` and `box`
3143 symbols, as appropriate. For example, these two matches on `x: &i32` are
3148 let y = match *x { 0 => "zero", _ => "some" };
3149 let z = match x { &0 => "zero", _ => "some" };
3154 A pattern that's just an identifier, like `Nil` in the previous example, could
3155 either refer to an enum variant that's in scope, or bind a new variable. The
3156 compiler resolves this ambiguity by forbidding variable bindings that occur in
3157 `match` patterns from shadowing names of variants that are in scope. For
3158 example, wherever `List` is in scope, a `match` pattern would not be able to
3159 bind `Nil` as a new name. The compiler interprets a variable pattern `x` as a
3160 binding _only_ if there is no variant named `x` in scope. A convention you can
3161 use to avoid conflicts is simply to name variants with upper-case letters, and
3162 local variables with lower-case letters.
3164 Multiple match patterns may be joined with the `|` operator. A range of values
3165 may be specified with `...`. For example:
3170 let message = match x {
3171 0 | 1 => "not many",
3177 Range patterns only work on scalar types (like integers and characters; not
3178 like arrays and structs, which have sub-components). A range pattern may not
3179 be a sub-range of another range pattern inside the same `match`.
3181 Finally, match patterns can accept *pattern guards* to further refine the
3182 criteria for matching a case. Pattern guards appear after the pattern and
3183 consist of a bool-typed expression following the `if` keyword. A pattern guard
3184 may refer to the variables bound within the pattern they follow.
3187 # let maybe_digit = Some(0);
3188 # fn process_digit(i: i32) { }
3189 # fn process_other(i: i32) { }
3191 let message = match maybe_digit {
3192 Some(x) if x < 10 => process_digit(x),
3193 Some(x) => process_other(x),
3198 ### If let expressions
3200 An `if let` expression is semantically identical to an `if` expression but in place
3201 of a condition expression it expects a refutable let statement. If the value of the
3202 expression on the right hand side of the let statement matches the pattern, the corresponding
3203 block will execute, otherwise flow proceeds to the first `else` block that follows.
3206 let dish = ("Ham", "Eggs");
3208 // this body will be skipped because the pattern is refuted
3209 if let ("Bacon", b) = dish {
3210 println!("Bacon is served with {}", b);
3213 // this body will execute
3214 if let ("Ham", b) = dish {
3215 println!("Ham is served with {}", b);
3221 A `while let` loop is semantically identical to a `while` loop but in place of a
3222 condition expression it expects a refutable let statement. If the value of the
3223 expression on the right hand side of the let statement matches the pattern, the
3224 loop body block executes and control returns to the pattern matching statement.
3225 Otherwise, the while expression completes.
3227 ### Return expressions
3229 Return expressions are denoted with the keyword `return`. Evaluating a `return`
3230 expression moves its argument into the designated output location for the
3231 current function call, destroys the current function activation frame, and
3232 transfers control to the caller frame.
3234 An example of a `return` expression:
3237 fn max(a: i32, b: i32) -> i32 {
3249 Every variable, item and value in a Rust program has a type. The _type_ of a
3250 *value* defines the interpretation of the memory holding it.
3252 Built-in types and type-constructors are tightly integrated into the language,
3253 in nontrivial ways that are not possible to emulate in user-defined types.
3254 User-defined types have limited capabilities.
3258 The primitive types are the following:
3260 * The boolean type `bool` with values `true` and `false`.
3261 * The machine types (integer and floating-point).
3262 * The machine-dependent integer types.
3266 The machine types are the following:
3268 * The unsigned word types `u8`, `u16`, `u32` and `u64`, with values drawn from
3269 the integer intervals [0, 2^8 - 1], [0, 2^16 - 1], [0, 2^32 - 1] and
3270 [0, 2^64 - 1] respectively.
3272 * The signed two's complement word types `i8`, `i16`, `i32` and `i64`, with
3273 values drawn from the integer intervals [-(2^(7)), 2^7 - 1],
3274 [-(2^(15)), 2^15 - 1], [-(2^(31)), 2^31 - 1], [-(2^(63)), 2^63 - 1]
3277 * The IEEE 754-2008 `binary32` and `binary64` floating-point types: `f32` and
3278 `f64`, respectively.
3280 #### Machine-dependent integer types
3282 The `usize` type is an unsigned integer type with the same number of bits as the
3283 platform's pointer type. It can represent every memory address in the process.
3285 The `isize` type is a signed integer type with the same number of bits as the
3286 platform's pointer type. The theoretical upper bound on object and array size
3287 is the maximum `isize` value. This ensures that `isize` can be used to calculate
3288 differences between pointers into an object or array and can address every byte
3289 within an object along with one byte past the end.
3293 The types `char` and `str` hold textual data.
3295 A value of type `char` is a [Unicode scalar value](
3296 http://www.unicode.org/glossary/#unicode_scalar_value) (i.e. a code point that
3297 is not a surrogate), represented as a 32-bit unsigned word in the 0x0000 to
3298 0xD7FF or 0xE000 to 0x10FFFF range. A `[char]` array is effectively an UCS-4 /
3301 A value of type `str` is a Unicode string, represented as an array of 8-bit
3302 unsigned bytes holding a sequence of UTF-8 code points. Since `str` is of
3303 unknown size, it is not a _first-class_ type, but can only be instantiated
3304 through a pointer type, such as `&str`.
3308 A tuple *type* is a heterogeneous product of other types, called the *elements*
3309 of the tuple. It has no nominal name and is instead structurally typed.
3311 Tuple types and values are denoted by listing the types or values of their
3312 elements, respectively, in a parenthesized, comma-separated list.
3314 Because tuple elements don't have a name, they can only be accessed by
3315 pattern-matching or by using `N` directly as a field to access the
3318 An example of a tuple type and its use:
3321 type Pair<'a> = (i32, &'a str);
3322 let p: Pair<'static> = (10, "hello");
3324 assert!(b != "world");
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'.
3336 * `&[T]`, a 'slice'.
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
3344 An example of each kind:
3347 let vec: Vec<i32> = vec![1, 2, 3];
3348 let arr: [i32; 3] = [1, 2, 3];
3349 let s: &[i32] = &vec[..];
3352 As you can see, the `vec!` macro allows you to create a `Vec<T>` easily. The
3353 `vec!` macro is also part of the standard library, rather than the language.
3355 All in-bounds elements of arrays, and slices are always initialized, and access
3356 to an array or slice is always bounds-checked.
3360 A `struct` *type* is a heterogeneous product of other types, called the
3361 *fields* of the type.[^structtype]
3363 [^structtype]: `struct` types are analogous to `struct` types in C,
3364 the *record* types of the ML family,
3365 or the *structure* types of the Lisp family.
3367 New instances of a `struct` can be constructed with a [struct
3368 expression](#structure-expressions).
3370 The memory layout of a `struct` is undefined by default to allow for compiler
3371 optimizations like field reordering, but it can be fixed with the
3372 `#[repr(...)]` attribute. In either case, fields may be given in any order in
3373 a corresponding struct *expression*; the resulting `struct` value will always
3374 have the same memory layout.
3376 The fields of a `struct` may be qualified by [visibility
3377 modifiers](#visibility-and-privacy), to allow access to data in a
3378 structure outside a module.
3380 A _tuple struct_ type is just like a structure type, except that the fields are
3383 A _unit-like struct_ type is like a structure type, except that it has no
3384 fields. The one value constructed by the associated [structure
3385 expression](#structure-expressions) is the only value that inhabits such a
3388 ### Enumerated types
3390 An *enumerated type* is a nominal, heterogeneous disjoint union type, denoted
3391 by the name of an [`enum` item](#enumerations). [^enumtype]
3393 [^enumtype]: The `enum` type is analogous to a `data` constructor declaration in
3394 ML, or a *pick ADT* in Limbo.
3396 An [`enum` item](#enumerations) declares both the type and a number of *variant
3397 constructors*, each of which is independently named and takes an optional tuple
3400 New instances of an `enum` can be constructed by calling one of the variant
3401 constructors, in a [call expression](#call-expressions).
3403 Any `enum` value consumes as much memory as the largest variant constructor for
3404 its corresponding `enum` type.
3406 Enum types cannot be denoted *structurally* as types, but must be denoted by
3407 named reference to an [`enum` item](#enumerations).
3411 Nominal types — [enumerations](#enumerated-types) and
3412 [structures](#structure-types) — may be recursive. That is, each `enum`
3413 constructor or `struct` field may refer, directly or indirectly, to the
3414 enclosing `enum` or `struct` type itself. Such recursion has restrictions:
3416 * Recursive types must include a nominal type in the recursion
3417 (not mere [type definitions](grammar.html#type-definitions),
3418 or other structural types such as [arrays](#array,-and-slice-types) or [tuples](#tuple-types)).
3419 * A recursive `enum` item must have at least one non-recursive constructor
3420 (in order to give the recursion a basis case).
3421 * The size of a recursive type must be finite;
3422 in other words the recursive fields of the type must be [pointer types](#pointer-types).
3423 * Recursive type definitions can cross module boundaries, but not module *visibility* boundaries,
3424 or crate boundaries (in order to simplify the module system and type checker).
3426 An example of a *recursive* type and its use:
3431 Cons(T, Box<List<T>>)
3434 let a: List<i32> = List::Cons(7, Box::new(List::Cons(13, Box::new(List::Nil))));
3439 All pointers in Rust are explicit first-class values. They can be copied,
3440 stored into data structures, and returned from functions. There are two
3441 varieties of pointer in Rust:
3444 : These point to memory _owned by some other value_.
3445 A reference type is written `&type`,
3446 or `&'a type` when you need to specify an explicit lifetime.
3447 Copying a reference is a "shallow" operation:
3448 it involves only copying the pointer itself.
3449 Releasing a reference has no effect on the value it points to,
3450 but a reference of a temporary value will keep it alive during the scope
3451 of the reference itself.
3453 * Raw pointers (`*`)
3454 : Raw pointers are pointers without safety or liveness guarantees.
3455 Raw pointers are written as `*const T` or `*mut T`,
3456 for example `*const i32` means a raw pointer to a 32-bit integer.
3457 Copying or dropping a raw pointer has no effect on the lifecycle of any
3458 other value. Dereferencing a raw pointer or converting it to any other
3459 pointer type is an [`unsafe` operation](#unsafe-functions).
3460 Raw pointers are generally discouraged in Rust code;
3461 they exist to support interoperability with foreign code,
3462 and writing performance-critical or low-level functions.
3464 The standard library contains additional 'smart pointer' types beyond references
3469 The function type constructor `fn` forms new function types. A function type
3470 consists of a possibly-empty set of function-type modifiers (such as `unsafe`
3471 or `extern`), a sequence of input types and an output type.
3473 An example of a `fn` type:
3476 fn add(x: i32, y: i32) -> i32 {
3480 let mut x = add(5,7);
3482 type Binop = fn(i32, i32) -> i32;
3483 let bo: Binop = add;
3487 #### Function types for specific items
3489 Internally to the compiler, there are also function types that are specific to a particular
3490 function item. In the following snippet, for example, the internal types of the functions
3491 `foo` and `bar` are different, despite the fact that they have the same signature:
3498 The types of `foo` and `bar` can both be implicitly coerced to the fn
3499 pointer type `fn()`. There is currently no syntax for unique fn types,
3500 though the compiler will emit a type like `fn() {foo}` in error
3501 messages to indicate "the unique fn type for the function `foo`".
3505 A [lambda expression](#lambda-expressions) produces a closure value with
3506 a unique, anonymous type that cannot be written out.
3508 Depending on the requirements of the closure, its type implements one or
3509 more of the closure traits:
3512 : The closure can be called once. A closure called as `FnOnce`
3513 can move out values from its environment.
3516 : The closure can be called multiple times as mutable. A closure called as
3517 `FnMut` can mutate values from its environment. `FnMut` implies
3521 : The closure can be called multiple times through a shared reference.
3522 A closure called as `Fn` can neither move out from nor mutate values
3523 from its environment. `Fn` implies `FnMut` and `FnOnce`.
3528 In Rust, a type like `&SomeTrait` or `Box<SomeTrait>` is called a _trait object_.
3529 Each instance of a trait object includes:
3531 - a pointer to an instance of a type `T` that implements `SomeTrait`
3532 - a _virtual method table_, often just called a _vtable_, which contains, for
3533 each method of `SomeTrait` that `T` implements, a pointer to `T`'s
3534 implementation (i.e. a function pointer).
3536 The purpose of trait objects is to permit "late binding" of methods. A call to
3537 a method on a trait object is only resolved to a vtable entry at compile time.
3538 The actual implementation for each vtable entry can vary on an object-by-object
3541 Note that for a trait object to be instantiated, the trait must be
3542 _object-safe_. Object safety rules are defined in [RFC 255].
3544 [RFC 255]: https://github.com/rust-lang/rfcs/blob/master/text/0255-object-safety.md
3546 Given a pointer-typed expression `E` of type `&T` or `Box<T>`, where `T`
3547 implements trait `R`, casting `E` to the corresponding pointer type `&R` or
3548 `Box<R>` results in a value of the _trait object_ `R`. This result is
3549 represented as a pair of pointers: the vtable pointer for the `T`
3550 implementation of `R`, and the pointer value of `E`.
3552 An example of a trait object:
3556 fn stringify(&self) -> String;
3559 impl Printable for i32 {
3560 fn stringify(&self) -> String { self.to_string() }
3563 fn print(a: Box<Printable>) {
3564 println!("{}", a.stringify());
3568 print(Box::new(10) as Box<Printable>);
3572 In this example, the trait `Printable` occurs as a trait object in both the
3573 type signature of `print`, and the cast expression in `main`.
3577 Within the body of an item that has type parameter declarations, the names of
3578 its type parameters are types:
3581 fn to_vec<A: Clone>(xs: &[A]) -> Vec<A> {
3585 let first: A = xs[0].clone();
3586 let mut rest: Vec<A> = to_vec(&xs[1..]);
3587 rest.insert(0, first);
3592 Here, `first` has type `A`, referring to `to_vec`'s `A` type parameter; and `rest`
3593 has type `Vec<A>`, a vector with element type `A`.
3597 The special type `Self` has a meaning within traits and impls. In a trait definition, it refers
3598 to an implicit type parameter representing the "implementing" type. In an impl,
3599 it is an alias for the implementing type. For example, in:
3603 fn make_string(&self) -> String;
3606 impl Printable for String {
3607 fn make_string(&self) -> String {
3613 The notation `&self` is a shorthand for `self: &Self`. In this case,
3614 in the impl, `Self` refers to the value of type `String` that is the
3615 receiver for a call to the method `make_string`.
3619 Subtyping is implicit and can occur at any stage in type checking or
3620 inference. Subtyping in Rust is very restricted and occurs only due to
3621 variance with respect to lifetimes and between types with higher ranked
3622 lifetimes. If we were to erase lifetimes from types, then the only subtyping
3623 would be due to type equality.
3625 Consider the following example: string literals always have `'static`
3626 lifetime. Nevertheless, we can assign `s` to `t`:
3630 let s: &'static str = "hi";
3634 Since `'static` "lives longer" than `'a`, `&'static str` is a subtype of
3639 Coercions are defined in [RFC401]. A coercion is implicit and has no syntax.
3641 [RFC401]: https://github.com/rust-lang/rfcs/blob/master/text/0401-coercions.md
3645 A coercion can only occur at certain coercion sites in a program; these are
3646 typically places where the desired type is explicit or can be dervied by
3647 propagation from explicit types (without type inference). Possible coercion
3650 * `let` statements where an explicit type is given.
3652 In `let _: U = e;`, `e` is coerced to have type `U`.
3654 * `static` and `const` statements (similar to `let` statements).
3656 * arguments for function calls.
3658 The value being coerced is the
3659 actual parameter and it is coerced to the type of the formal parameter. For
3660 example, let `foo` be defined as `fn foo(x: U) { ... }` and call it as
3661 `foo(e);`. Then `e` is coerced to have type `U`;
3663 * instantiations of struct or variant fields.
3665 Assume we have a `struct
3666 Foo { x: U }` and instantiate it as `Foo { x: e }`. Then `e` is coerced to
3669 * function results (either the final line of a block if it is not semicolon
3670 terminated or any expression in a `return` statement).
3672 In `fn foo() -> U { e }`, `e` is coerced to to have type `U`.
3674 If the expression in one of these coercion sites is a coercion-propagating
3675 expression, then the relevant sub-expressions in that expression are also
3676 coercion sites. Propagation recurses from these new coercion sites.
3677 Propagating expressions and their relevant sub-expressions are:
3679 * array literals, where the array has type `[U; n]`. Each sub-expression in
3680 the array literal is a coercion site for coercion to type `U`.
3682 * array literals with repeating syntax, where the array has type `[U; n]`. The
3683 repeated sub-expression is a coercion site for coercion to type `U`.
3685 * tuples, where a tuple is a coercion site to type `(U_0, U_1, ..., U_n)`.
3686 Each sub-expression is a coercion site to the respective type, e.g. the
3687 zeroth sub-expression is a coercion site to type `U_0`.
3689 * parenthesised sub-expressions (`(e)`). If the expression has type `U`, then
3690 the sub-expression is a coercion site to `U`.
3692 * blocks. If a block has type `U`, then the last expression in the block (if
3693 it is not semicolon-terminated) is a coercion site to `U`. This includes
3694 blocks which are part of control flow statements, such as `if`/`else`, if
3695 the block has a known type.
3699 Coercion is allowed between the following types:
3701 * `T` to `U` if `T` is a subtype of `U` (*reflexive case*).
3703 * `T_1` to `T_3` where `T_1` coerces to `T_2` and `T_2` coerces to `T_3`
3704 (*transitive case*).
3706 Note that this is not fully supported yet
3710 * `*mut T` to `*const T`.
3712 * `&T` to `*const T`.
3714 * `&mut T` to `*mut T`.
3716 * `&T` to `&U` if `T` implements `Deref<Target = U>`. For example:
3719 use std::ops::Deref;
3721 struct CharContainer {
3725 impl Deref for CharContainer {
3728 fn deref<'a>(&'a self) -> &'a char {
3733 fn foo(arg: &char) {}
3736 let x = &mut CharContainer { value: 'y' };
3737 foo(x); //&mut CharContainer is coerced to &char.
3740 * `&mut T` to `&mut U` if `T` implements `DerefMut<Target = U>`.
3742 * TyCtor(`T`) to TyCtor(coerce_inner(`T`)), where TyCtor(`T`) is one of
3750 - coerce_inner(`[T, ..n]`) = `[T]`
3751 - coerce_inner(`T`) = `U` where `T` is a concrete type which implements the
3754 In the future, coerce_inner will be recursively extended to tuples and
3755 structs. In addition, coercions from sub-traits to super-traits will be
3756 added. See [RFC401] for more details.
3760 Several traits define special evaluation behavior.
3764 The `Copy` trait changes the semantics of a type implementing it. Values whose
3765 type implements `Copy` are copied rather than moved upon assignment.
3767 ## The `Sized` trait
3769 The `Sized` trait indicates that the size of this type is known at compile-time.
3773 The `Drop` trait provides a destructor, to be run whenever a value of this type
3776 ## The `Deref` trait
3778 The `Deref<Target = U>` trait allows a type to implicitly implement all the methods
3779 of the type `U`. When attempting to resolve a method call, the compiler will search
3780 the top-level type for the implementation of the called method. If no such method is
3781 found, `.deref()` is called and the compiler continues to search for the method
3782 implementation in the returned type `U`.
3786 A Rust program's memory consists of a static set of *items* and a *heap*.
3787 Immutable portions of the heap may be safely shared between threads, mutable
3788 portions may not be safely shared, but several mechanisms for effectively-safe
3789 sharing of mutable values, built on unsafe code but enforcing a safe locking
3790 discipline, exist in the standard library.
3792 Allocations in the stack consist of *variables*, and allocations in the heap
3795 ### Memory allocation and lifetime
3797 The _items_ of a program are those functions, modules and types that have their
3798 value calculated at compile-time and stored uniquely in the memory image of the
3799 rust process. Items are neither dynamically allocated nor freed.
3801 The _heap_ is a general term that describes boxes. The lifetime of an
3802 allocation in the heap depends on the lifetime of the box values pointing to
3803 it. Since box values may themselves be passed in and out of frames, or stored
3804 in the heap, heap allocations may outlive the frame they are allocated within.
3806 ### Memory ownership
3808 When a stack frame is exited, its local allocations are all released, and its
3809 references to boxes are dropped.
3813 A _variable_ is a component of a stack frame, either a named function parameter,
3814 an anonymous [temporary](#lvalues,-rvalues-and-temporaries), or a named local
3817 A _local variable_ (or *stack-local* allocation) holds a value directly,
3818 allocated within the stack's memory. The value is a part of the stack frame.
3820 Local variables are immutable unless declared otherwise like: `let mut x = ...`.
3822 Function parameters are immutable unless declared with `mut`. The `mut` keyword
3823 applies only to the following parameter (so `|mut x, y|` and `fn f(mut x:
3824 Box<i32>, y: Box<i32>)` declare one mutable variable `x` and one immutable
3827 Methods that take either `self` or `Box<Self>` can optionally place them in a
3828 mutable variable by prefixing them with `mut` (similar to regular arguments):
3832 fn change(mut self) -> Self;
3833 fn modify(mut self: Box<Self>) -> Box<Self>;
3837 Local variables are not initialized when allocated; the entire frame worth of
3838 local variables are allocated at once, on frame-entry, in an uninitialized
3839 state. Subsequent statements within a function may or may not initialize the
3840 local variables. Local variables can be used only after they have been
3841 initialized; this is enforced by the compiler.
3845 The Rust compiler supports various methods to link crates together both
3846 statically and dynamically. This section will explore the various methods to
3847 link Rust crates together, and more information about native libraries can be
3848 found in the [ffi section of the book][ffi].
3850 In one session of compilation, the compiler can generate multiple artifacts
3851 through the usage of either command line flags or the `crate_type` attribute.
3852 If one or more command line flag is specified, all `crate_type` attributes will
3853 be ignored in favor of only building the artifacts specified by command line.
3855 * `--crate-type=bin`, `#[crate_type = "bin"]` - A runnable executable will be
3856 produced. This requires that there is a `main` function in the crate which
3857 will be run when the program begins executing. This will link in all Rust and
3858 native dependencies, producing a distributable binary.
3860 * `--crate-type=lib`, `#[crate_type = "lib"]` - A Rust library will be produced.
3861 This is an ambiguous concept as to what exactly is produced because a library
3862 can manifest itself in several forms. The purpose of this generic `lib` option
3863 is to generate the "compiler recommended" style of library. The output library
3864 will always be usable by rustc, but the actual type of library may change from
3865 time-to-time. The remaining output types are all different flavors of
3866 libraries, and the `lib` type can be seen as an alias for one of them (but the
3867 actual one is compiler-defined).
3869 * `--crate-type=dylib`, `#[crate_type = "dylib"]` - A dynamic Rust library will
3870 be produced. This is different from the `lib` output type in that this forces
3871 dynamic library generation. The resulting dynamic library can be used as a
3872 dependency for other libraries and/or executables. This output type will
3873 create `*.so` files on linux, `*.dylib` files on osx, and `*.dll` files on
3876 * `--crate-type=staticlib`, `#[crate_type = "staticlib"]` - A static system
3877 library will be produced. This is different from other library outputs in that
3878 the Rust compiler will never attempt to link to `staticlib` outputs. The
3879 purpose of this output type is to create a static library containing all of
3880 the local crate's code along with all upstream dependencies. The static
3881 library is actually a `*.a` archive on linux and osx and a `*.lib` file on
3882 windows. This format is recommended for use in situations such as linking
3883 Rust code into an existing non-Rust application because it will not have
3884 dynamic dependencies on other Rust code.
3886 * `--crate-type=rlib`, `#[crate_type = "rlib"]` - A "Rust library" file will be
3887 produced. This is used as an intermediate artifact and can be thought of as a
3888 "static Rust library". These `rlib` files, unlike `staticlib` files, are
3889 interpreted by the Rust compiler in future linkage. This essentially means
3890 that `rustc` will look for metadata in `rlib` files like it looks for metadata
3891 in dynamic libraries. This form of output is used to produce statically linked
3892 executables as well as `staticlib` outputs.
3894 Note that these outputs are stackable in the sense that if multiple are
3895 specified, then the compiler will produce each form of output at once without
3896 having to recompile. However, this only applies for outputs specified by the
3897 same method. If only `crate_type` attributes are specified, then they will all
3898 be built, but if one or more `--crate-type` command line flag is specified,
3899 then only those outputs will be built.
3901 With all these different kinds of outputs, if crate A depends on crate B, then
3902 the compiler could find B in various different forms throughout the system. The
3903 only forms looked for by the compiler, however, are the `rlib` format and the
3904 dynamic library format. With these two options for a dependent library, the
3905 compiler must at some point make a choice between these two formats. With this
3906 in mind, the compiler follows these rules when determining what format of
3907 dependencies will be used:
3909 1. If a static library is being produced, all upstream dependencies are
3910 required to be available in `rlib` formats. This requirement stems from the
3911 reason that a dynamic library cannot be converted into a static format.
3913 Note that it is impossible to link in native dynamic dependencies to a static
3914 library, and in this case warnings will be printed about all unlinked native
3915 dynamic dependencies.
3917 2. If an `rlib` file is being produced, then there are no restrictions on what
3918 format the upstream dependencies are available in. It is simply required that
3919 all upstream dependencies be available for reading metadata from.
3921 The reason for this is that `rlib` files do not contain any of their upstream
3922 dependencies. It wouldn't be very efficient for all `rlib` files to contain a
3923 copy of `libstd.rlib`!
3925 3. If an executable is being produced and the `-C prefer-dynamic` flag is not
3926 specified, then dependencies are first attempted to be found in the `rlib`
3927 format. If some dependencies are not available in an rlib format, then
3928 dynamic linking is attempted (see below).
3930 4. If a dynamic library or an executable that is being dynamically linked is
3931 being produced, then the compiler will attempt to reconcile the available
3932 dependencies in either the rlib or dylib format to create a final product.
3934 A major goal of the compiler is to ensure that a library never appears more
3935 than once in any artifact. For example, if dynamic libraries B and C were
3936 each statically linked to library A, then a crate could not link to B and C
3937 together because there would be two copies of A. The compiler allows mixing
3938 the rlib and dylib formats, but this restriction must be satisfied.
3940 The compiler currently implements no method of hinting what format a library
3941 should be linked with. When dynamically linking, the compiler will attempt to
3942 maximize dynamic dependencies while still allowing some dependencies to be
3943 linked in via an rlib.
3945 For most situations, having all libraries available as a dylib is recommended
3946 if dynamically linking. For other situations, the compiler will emit a
3947 warning if it is unable to determine which formats to link each library with.
3949 In general, `--crate-type=bin` or `--crate-type=lib` should be sufficient for
3950 all compilation needs, and the other options are just available if more
3951 fine-grained control is desired over the output format of a Rust crate.
3953 # Appendix: Rationales and design tradeoffs
3957 # Appendix: Influences
3959 Rust is not a particularly original language, with design elements coming from
3960 a wide range of sources. Some of these are listed below (including elements
3961 that have since been removed):
3963 * SML, OCaml: algebraic datatypes, pattern matching, type inference,
3964 semicolon statement separation
3965 * C++: references, RAII, smart pointers, move semantics, monomorphisation,
3967 * ML Kit, Cyclone: region based memory management
3968 * Haskell (GHC): typeclasses, type families
3969 * Newsqueak, Alef, Limbo: channels, concurrency
3970 * Erlang: message passing, thread failure, ~~linked thread failure~~,
3971 ~~lightweight concurrency~~
3972 * Swift: optional bindings
3973 * Scheme: hygienic macros
3975 * Ruby: ~~block syntax~~
3976 * NIL, Hermes: ~~typestate~~
3977 * [Unicode Annex #31](http://www.unicode.org/reports/tr31/): identifier and
3980 [ffi]: book/ffi.html
3981 [plugin]: book/compiler-plugins.html