5 This document is the primary reference for the Rust programming language. It
6 provides three kinds of material:
8 - Chapters that informally describe each language construct and their use.
9 - Chapters that informally describe the memory model, concurrency model,
10 runtime services, linkage model and debugging facilities.
11 - Appendix chapters providing rationale and references to languages that
12 influenced the design.
14 This document does not serve as an introduction to the language. Background
15 familiarity with the language is assumed. A separate [book] is available to
16 help acquire such background familiarity.
18 This document also does not serve as a reference to the [standard] library
19 included in the language distribution. Those libraries are documented
20 separately by extracting documentation attributes from their source code. Many
21 of the features that one might expect to be language features are library
22 features in Rust, so what you're looking for may be there, not here.
24 You may also be interested in the [grammar].
26 [book]: book/index.html
27 [standard]: std/index.html
28 [grammar]: grammar.html
32 ## Unicode productions
34 A few productions in Rust's grammar permit Unicode code points outside the
35 ASCII range. We define these productions in terms of character properties
36 specified in the Unicode standard, rather than in terms of ASCII-range code
37 points. The grammar has a [Special Unicode Productions][unicodeproductions]
38 section that lists these productions.
40 [unicodeproductions]: grammar.html#special-unicode-productions
42 ## String table productions
44 Some rules in the grammar — notably [unary
45 operators](#unary-operator-expressions), [binary
46 operators](#binary-operator-expressions), and [keywords][keywords] — are
47 given in a simplified form: as a listing of a table of unquoted, printable
48 whitespace-separated strings. These cases form a subset of the rules regarding
49 the [token](#tokens) rule, and are assumed to be the result of a
50 lexical-analysis phase feeding the parser, driven by a DFA, operating over the
51 disjunction of all such string table entries.
53 [keywords]: grammar.html#keywords
55 When such a string enclosed in double-quotes (`"`) occurs inside the grammar,
56 it is an implicit reference to a single member of such a string table
57 production. See [tokens](#tokens) for more information.
63 Rust input is interpreted as a sequence of Unicode code points encoded in UTF-8.
64 Most Rust grammar rules are defined in terms of printable ASCII-range
65 code points, but a small number are defined in terms of Unicode properties or
66 explicit code point lists. [^inputformat]
68 [^inputformat]: Substitute definitions for the special Unicode productions are
69 provided to the grammar verifier, restricted to ASCII range, when verifying the
70 grammar in this document.
74 An identifier is any nonempty Unicode[^non_ascii_idents] string of the following form:
76 [^non_ascii_idents]: Non-ASCII characters in identifiers are currently feature
77 gated. This is expected to improve soon.
81 * The first character has property `XID_start`
82 * The remaining characters have property `XID_continue`
86 * The first character is `_`
87 * The identifier is more than one character, `_` alone is not an identifier
88 * The remaining characters have property `XID_continue`
90 that does _not_ occur in the set of [keywords][keywords].
92 > **Note**: `XID_start` and `XID_continue` as character properties cover the
93 > character ranges used to form the more familiar C and Java language-family
98 Comments in Rust code follow the general C++ style of line (`//`) and
99 block (`/* ... */`) comment forms. Nested block comments are supported.
101 Line comments beginning with exactly _three_ slashes (`///`), and block
102 comments beginning with exactly one repeated asterisk in the block-open
103 sequence (`/**`), are interpreted as a special syntax for `doc`
104 [attributes](#attributes). That is, they are equivalent to writing
105 `#[doc="..."]` around the body of the comment, i.e., `/// Foo` turns into
108 Line comments beginning with `//!` and block comments beginning with `/*!` are
109 doc comments that apply to the parent of the comment, rather than the item
110 that follows. That is, they are equivalent to writing `#![doc="..."]` around
111 the body of the comment. `//!` comments are usually used to document
112 modules that occupy a source file.
114 Non-doc comments are interpreted as a form of whitespace.
118 Whitespace is any non-empty string containing only the following characters:
120 - `U+0020` (space, `' '`)
121 - `U+0009` (tab, `'\t'`)
122 - `U+000A` (LF, `'\n'`)
123 - `U+000D` (CR, `'\r'`)
125 Rust is a "free-form" language, meaning that all forms of whitespace serve only
126 to separate _tokens_ in the grammar, and have no semantic significance.
128 A Rust program has identical meaning if each whitespace element is replaced
129 with any other legal whitespace element, such as a single space character.
133 Tokens are primitive productions in the grammar defined by regular
134 (non-recursive) languages. "Simple" tokens are given in [string table
135 production](#string-table-productions) form, and occur in the rest of the
136 grammar as double-quoted strings. Other tokens have exact rules given.
140 A literal is an expression consisting of a single token, rather than a sequence
141 of tokens, that immediately and directly denotes the value it evaluates to,
142 rather than referring to it by name or some other evaluation rule. A literal is
143 a form of constant expression, so is evaluated (primarily) at compile time.
147 ##### Characters and strings
149 | | Example | `#` sets | Characters | Escapes |
150 |----------------------------------------------|-----------------|------------|-------------|---------------------|
151 | [Character](#character-literals) | `'H'` | `N/A` | All Unicode | `\'` & [Byte](#byte-escapes) & [Unicode](#unicode-escapes) |
152 | [String](#string-literals) | `"hello"` | `N/A` | All Unicode | `\"` & [Byte](#byte-escapes) & [Unicode](#unicode-escapes) |
153 | [Raw](#raw-string-literals) | `r#"hello"#` | `0...` | All Unicode | `N/A` |
154 | [Byte](#byte-literals) | `b'H'` | `N/A` | All ASCII | `\'` & [Byte](#byte-escapes) |
155 | [Byte string](#byte-string-literals) | `b"hello"` | `N/A` | All ASCII | `\"` & [Byte](#byte-escapes) |
156 | [Raw byte string](#raw-byte-string-literals) | `br#"hello"#` | `0...` | All ASCII | `N/A` |
162 | `\x7F` | 8-bit character code (exactly 2 digits) |
164 | `\r` | Carriage return |
168 ##### Unicode escapes
171 | `\u{7FFF}` | 24-bit Unicode character code (up to 6 digits) |
175 | [Number literals](#number-literals)`*` | Example | Exponentiation | Suffixes |
176 |----------------------------------------|---------|----------------|----------|
177 | Decimal integer | `98_222` | `N/A` | Integer suffixes |
178 | Hex integer | `0xff` | `N/A` | Integer suffixes |
179 | Octal integer | `0o77` | `N/A` | Integer suffixes |
180 | Binary integer | `0b1111_0000` | `N/A` | Integer suffixes |
181 | Floating-point | `123.0E+77` | `Optional` | Floating-point suffixes |
183 `*` All number literals allow `_` as a visual separator: `1_234.0E+18f64`
186 | Integer | Floating-point |
187 |---------|----------------|
188 | `u8`, `i8`, `u16`, `i16`, `u32`, `i32`, `u64`, `i64`, `isize`, `usize` | `f32`, `f64` |
190 #### Character and string literals
192 ##### Character literals
194 A _character literal_ is a single Unicode character enclosed within two
195 `U+0027` (single-quote) characters, with the exception of `U+0027` itself,
196 which must be _escaped_ by a preceding `U+005C` character (`\`).
198 ##### String literals
200 A _string literal_ is a sequence of any Unicode characters enclosed within two
201 `U+0022` (double-quote) characters, with the exception of `U+0022` itself,
202 which must be _escaped_ by a preceding `U+005C` character (`\`).
204 Line-break characters are allowed in string literals. Normally they represent
205 themselves (i.e. no translation), but as a special exception, when a `U+005C`
206 character (`\`) occurs immediately before the newline, the `U+005C` character,
207 the newline, and all whitespace at the beginning of the next line are ignored.
208 Thus `a` and `b` are equal:
218 ##### Character escapes
220 Some additional _escapes_ are available in either character or non-raw string
221 literals. An escape starts with a `U+005C` (`\`) and continues with one of the
224 * An _8-bit code point escape_ starts with `U+0078` (`x`) and is
225 followed by exactly two _hex digits_. It denotes the Unicode code point
226 equal to the provided hex value.
227 * A _24-bit code point escape_ starts with `U+0075` (`u`) and is followed
228 by up to six _hex digits_ surrounded by braces `U+007B` (`{`) and `U+007D`
229 (`}`). It denotes the Unicode code point equal to the provided hex value.
230 * A _whitespace escape_ is one of the characters `U+006E` (`n`), `U+0072`
231 (`r`), or `U+0074` (`t`), denoting the Unicode values `U+000A` (LF),
232 `U+000D` (CR) or `U+0009` (HT) respectively.
233 * The _backslash escape_ is the character `U+005C` (`\`) which must be
234 escaped in order to denote *itself*.
236 ##### Raw string literals
238 Raw string literals do not process any escapes. They start with the character
239 `U+0072` (`r`), followed by zero or more of the character `U+0023` (`#`) and a
240 `U+0022` (double-quote) character. The _raw string body_ can contain any sequence
241 of Unicode characters and is terminated only by another `U+0022` (double-quote)
242 character, followed by the same number of `U+0023` (`#`) characters that preceded
243 the opening `U+0022` (double-quote) character.
245 All Unicode characters contained in the raw string body represent themselves,
246 the characters `U+0022` (double-quote) (except when followed by at least as
247 many `U+0023` (`#`) characters as were used to start the raw string literal) or
248 `U+005C` (`\`) do not have any special meaning.
250 Examples for string literals:
253 "foo"; r"foo"; // foo
254 "\"foo\""; r#""foo""#; // "foo"
257 r##"foo #"# bar"##; // foo #"# bar
259 "\x52"; "R"; r"R"; // R
260 "\\x52"; r"\x52"; // \x52
263 #### Byte and byte string literals
267 A _byte literal_ is a single ASCII character (in the `U+0000` to `U+007F`
268 range) or a single _escape_ preceded by the characters `U+0062` (`b`) and
269 `U+0027` (single-quote), and followed by the character `U+0027`. If the character
270 `U+0027` is present within the literal, it must be _escaped_ by a preceding
271 `U+005C` (`\`) character. It is equivalent to a `u8` unsigned 8-bit integer
274 ##### Byte string literals
276 A non-raw _byte string literal_ is a sequence of ASCII characters and _escapes_,
277 preceded by the characters `U+0062` (`b`) and `U+0022` (double-quote), and
278 followed by the character `U+0022`. If the character `U+0022` is present within
279 the literal, it must be _escaped_ by a preceding `U+005C` (`\`) character.
280 Alternatively, a byte string literal can be a _raw byte string literal_, defined
281 below. A byte string literal of length `n` is equivalent to a `&'static [u8; n]` borrowed fixed-sized array
282 of unsigned 8-bit integers.
284 Some additional _escapes_ are available in either byte or non-raw byte string
285 literals. An escape starts with a `U+005C` (`\`) and continues with one of the
288 * A _byte escape_ escape starts with `U+0078` (`x`) and is
289 followed by exactly two _hex digits_. It denotes the byte
290 equal to the provided hex value.
291 * A _whitespace escape_ is one of the characters `U+006E` (`n`), `U+0072`
292 (`r`), or `U+0074` (`t`), denoting the bytes values `0x0A` (ASCII LF),
293 `0x0D` (ASCII CR) or `0x09` (ASCII HT) respectively.
294 * The _backslash escape_ is the character `U+005C` (`\`) which must be
295 escaped in order to denote its ASCII encoding `0x5C`.
297 ##### Raw byte string literals
299 Raw byte string literals do not process any escapes. They start with the
300 character `U+0062` (`b`), followed by `U+0072` (`r`), followed by zero or more
301 of the character `U+0023` (`#`), and a `U+0022` (double-quote) character. The
302 _raw string body_ can contain any sequence of ASCII characters and is terminated
303 only by another `U+0022` (double-quote) character, followed by the same number of
304 `U+0023` (`#`) characters that preceded the opening `U+0022` (double-quote)
305 character. A raw byte string literal can not contain any non-ASCII byte.
307 All characters contained in the raw string body represent their ASCII encoding,
308 the characters `U+0022` (double-quote) (except when followed by at least as
309 many `U+0023` (`#`) characters as were used to start the raw string literal) or
310 `U+005C` (`\`) do not have any special meaning.
312 Examples for byte string literals:
315 b"foo"; br"foo"; // foo
316 b"\"foo\""; br#""foo""#; // "foo"
319 br##"foo #"# bar"##; // foo #"# bar
321 b"\x52"; b"R"; br"R"; // R
322 b"\\x52"; br"\x52"; // \x52
327 A _number literal_ is either an _integer literal_ or a _floating-point
328 literal_. The grammar for recognizing the two kinds of literals is mixed.
330 ##### Integer literals
332 An _integer literal_ has one of four forms:
334 * A _decimal literal_ starts with a *decimal digit* and continues with any
335 mixture of *decimal digits* and _underscores_.
336 * A _hex literal_ starts with the character sequence `U+0030` `U+0078`
337 (`0x`) and continues as any mixture of hex digits and underscores.
338 * An _octal literal_ starts with the character sequence `U+0030` `U+006F`
339 (`0o`) and continues as any mixture of octal digits and underscores.
340 * A _binary literal_ starts with the character sequence `U+0030` `U+0062`
341 (`0b`) and continues as any mixture of binary digits and underscores.
343 Like any literal, an integer literal may be followed (immediately,
344 without any spaces) by an _integer suffix_, which forcibly sets the
345 type of the literal. The integer suffix must be the name of one of the
346 integral types: `u8`, `i8`, `u16`, `i16`, `u32`, `i32`, `u64`, `i64`,
349 The type of an _unsuffixed_ integer literal is determined by type inference:
351 * If an integer type can be _uniquely_ determined from the surrounding
352 program context, the unsuffixed integer literal has that type.
354 * If the program context under-constrains the type, it defaults to the
355 signed 32-bit integer `i32`.
357 * If the program context over-constrains the type, it is considered a
360 Examples of integer literals of various forms:
367 0o70_i16; // type i16
368 0b1111_1111_1001_0000_i32; // type i32
369 0usize; // type usize
372 ##### Floating-point literals
374 A _floating-point literal_ has one of two forms:
376 * A _decimal literal_ followed by a period character `U+002E` (`.`). This is
377 optionally followed by another decimal literal, with an optional _exponent_.
378 * A single _decimal literal_ followed by an _exponent_.
380 Like integer literals, a floating-point literal may be followed by a
381 suffix, so long as the pre-suffix part does not end with `U+002E` (`.`).
382 The suffix forcibly sets the type of the literal. There are two valid
383 _floating-point suffixes_, `f32` and `f64` (the 32-bit and 64-bit floating point
384 types), which explicitly determine the type of the literal.
386 The type of an _unsuffixed_ floating-point literal is determined by
389 * If a floating-point type can be _uniquely_ determined from the
390 surrounding program context, the unsuffixed floating-point literal
393 * If the program context under-constrains the type, it defaults to `f64`.
395 * If the program context over-constrains the type, it is considered a
398 Examples of floating-point literals of various forms:
401 123.0f64; // type f64
404 12E+99_f64; // type f64
405 let x: f64 = 2.; // type f64
408 This last example is different because it is not possible to use the suffix
409 syntax with a floating point literal ending in a period. `2.f64` would attempt
410 to call a method named `f64` on `2`.
412 The representation semantics of floating-point numbers are described in
413 ["Machine Types"](#machine-types).
415 #### Boolean literals
417 The two values of the boolean type are written `true` and `false`.
421 Symbols are a general class of printable [tokens](#tokens) that play structural
422 roles in a variety of grammar productions. They are a
423 set of remaining miscellaneous printable tokens that do not
424 otherwise appear as [unary operators](#unary-operator-expressions), [binary
425 operators](#binary-operator-expressions), or [keywords][keywords].
426 They are catalogued in [the Symbols section][symbols] of the Grammar document.
428 [symbols]: grammar.html#symbols
433 A _path_ is a sequence of one or more path components _logically_ separated by
434 a namespace qualifier (`::`). If a path consists of only one component, it may
435 refer to either an [item](#items) or a [variable](#variables) in a local control
436 scope. If a path has multiple components, it refers to an item.
438 Every item has a _canonical path_ within its crate, but the path naming an item
439 is only meaningful within a given crate. There is no global namespace across
440 crates; an item's canonical path merely identifies it within the crate.
442 Two examples of simple paths consisting of only identifier components:
449 Path components are usually [identifiers](#identifiers), but they may
450 also include angle-bracket-enclosed lists of type arguments. In
451 [expression](#expressions) context, the type argument list is given
452 after a `::` namespace qualifier in order to disambiguate it from a
453 relational expression involving the less-than symbol (`<`). In type
454 expression context, the final namespace qualifier is omitted.
456 Two examples of paths with type arguments:
459 # struct HashMap<K, V>(K,V);
461 # fn id<T>(t: T) -> T { t }
462 type T = HashMap<i32,String>; // Type arguments used in a type expression
463 let x = id::<i32>(10); // Type arguments used in a call expression
467 Paths can be denoted with various leading qualifiers to change the meaning of
470 * Paths starting with `::` are considered to be global paths where the
471 components of the path start being resolved from the crate root. Each
472 identifier in the path must resolve to an item.
480 ::a::foo(); // call a's foo function
486 * Paths starting with the keyword `super` begin resolution relative to the
487 parent module. Each further identifier must resolve to an item.
495 super::a::foo(); // call a's foo function
501 * Paths starting with the keyword `self` begin resolution relative to the
502 current module. Each further identifier must resolve to an item.
514 A number of minor features of Rust are not central enough to have their own
515 syntax, and yet are not implementable as functions. Instead, they are given
516 names, and invoked through a consistent syntax: `some_extension!(...)`.
518 Users of `rustc` can define new syntax extensions in two ways:
520 * [Compiler plugins][plugin] can include arbitrary Rust code that
521 manipulates syntax trees at compile time. Note that the interface
522 for compiler plugins is considered highly unstable.
524 * [Macros](book/macros.html) define new syntax in a higher-level,
529 `macro_rules` allows users to define syntax extension in a declarative way. We
530 call such extensions "macros by example" or simply "macros" — to be distinguished
531 from the "procedural macros" defined in [compiler plugins][plugin].
533 Currently, macros can expand to expressions, statements, items, or patterns.
535 (A `sep_token` is any token other than `*` and `+`. A `non_special_token` is
536 any token other than a delimiter or `$`.)
538 The macro expander looks up macro invocations by name, and tries each macro
539 rule in turn. It transcribes the first successful match. Matching and
540 transcription are closely related to each other, and we will describe them
545 The macro expander matches and transcribes every token that does not begin with
546 a `$` literally, including delimiters. For parsing reasons, delimiters must be
547 balanced, but they are otherwise not special.
549 In the matcher, `$` _name_ `:` _designator_ matches the nonterminal in the Rust
550 syntax named by _designator_. Valid designators are `item`, `block`, `stmt`,
551 `pat`, `expr`, `ty` (type), `ident`, `path`, `tt` (either side of the `=>`
552 in macro rules), and `meta` (contents of an attribute). In the transcriber, the
553 designator is already known, and so only the name of a matched nonterminal comes
554 after the dollar sign.
556 In both the matcher and transcriber, the Kleene star-like operator indicates
557 repetition. The Kleene star operator consists of `$` and parentheses, optionally
558 followed by a separator token, followed by `*` or `+`. `*` means zero or more
559 repetitions, `+` means at least one repetition. The parentheses are not matched or
560 transcribed. On the matcher side, a name is bound to _all_ of the names it
561 matches, in a structure that mimics the structure of the repetition encountered
562 on a successful match. The job of the transcriber is to sort that structure
565 The rules for transcription of these repetitions are called "Macro By Example".
566 Essentially, one "layer" of repetition is discharged at a time, and all of them
567 must be discharged by the time a name is transcribed. Therefore, `( $( $i:ident
568 ),* ) => ( $i )` is an invalid macro, but `( $( $i:ident ),* ) => ( $( $i:ident
569 ),* )` is acceptable (if trivial).
571 When Macro By Example encounters a repetition, it examines all of the `$`
572 _name_ s that occur in its body. At the "current layer", they all must repeat
573 the same number of times, so ` ( $( $i:ident ),* ; $( $j:ident ),* ) => ( $(
574 ($i,$j) ),* )` is valid if given the argument `(a,b,c ; d,e,f)`, but not
575 `(a,b,c ; d,e)`. The repetition walks through the choices at that layer in
576 lockstep, so the former input transcribes to `(a,d), (b,e), (c,f)`.
578 Nested repetitions are allowed.
580 ### Parsing limitations
582 The parser used by the macro system is reasonably powerful, but the parsing of
583 Rust syntax is restricted in two ways:
585 1. Macro definitions are required to include suitable separators after parsing
586 expressions and other bits of the Rust grammar. This implies that
587 a macro definition like `$i:expr [ , ]` is not legal, because `[` could be part
588 of an expression. A macro definition like `$i:expr,` or `$i:expr;` would be legal,
589 however, because `,` and `;` are legal separators. See [RFC 550] for more information.
590 2. The parser must have eliminated all ambiguity by the time it reaches a `$`
591 _name_ `:` _designator_. This requirement most often affects name-designator
592 pairs when they occur at the beginning of, or immediately after, a `$(...)*`;
593 requiring a distinctive token in front can solve the problem.
595 [RFC 550]: https://github.com/rust-lang/rfcs/blob/master/text/0550-macro-future-proofing.md
597 # Crates and source files
599 Although Rust, like any other language, can be implemented by an interpreter as
600 well as a compiler, the only existing implementation is a compiler,
602 always been designed to be compiled. For these reasons, this section assumes a
605 Rust's semantics obey a *phase distinction* between compile-time and
606 run-time.[^phase-distinction] Semantic rules that have a *static
607 interpretation* govern the success or failure of compilation, while
609 that have a *dynamic interpretation* govern the behavior of the program at
612 [^phase-distinction]: This distinction would also exist in an interpreter.
613 Static checks like syntactic analysis, type checking, and lints should
614 happen before the program is executed regardless of when it is executed.
616 The compilation model centers on artifacts called _crates_. Each compilation
617 processes a single crate in source form, and if successful, produces a single
618 crate in binary form: either an executable or some sort of
619 library.[^cratesourcefile]
621 [^cratesourcefile]: A crate is somewhat analogous to an *assembly* in the
622 ECMA-335 CLI model, a *library* in the SML/NJ Compilation Manager, a *unit*
623 in the Owens and Flatt module system, or a *configuration* in Mesa.
625 A _crate_ is a unit of compilation and linking, as well as versioning,
626 distribution and runtime loading. A crate contains a _tree_ of nested
627 [module](#modules) scopes. The top level of this tree is a module that is
628 anonymous (from the point of view of paths within the module) and any item
629 within a crate has a canonical [module path](#paths) denoting its location
630 within the crate's module tree.
632 The Rust compiler is always invoked with a single source file as input, and
633 always produces a single output crate. The processing of that source file may
634 result in other source files being loaded as modules. Source files have the
637 A Rust source file describes a module, the name and location of which —
638 in the module tree of the current crate — are defined from outside the
639 source file: either by an explicit `mod_item` in a referencing source file, or
640 by the name of the crate itself. Every source file is a module, but not every
641 module needs its own source file: [module definitions](#modules) can be nested
644 Each source file contains a sequence of zero or more `item` definitions, and
645 may optionally begin with any number of [attributes](#items-and-attributes)
646 that apply to the containing module, most of which influence the behavior of
647 the compiler. The anonymous crate module can have additional attributes that
648 apply to the crate as a whole.
651 // Specify the crate name.
652 #![crate_name = "projx"]
654 // Specify the type of output artifact.
655 #![crate_type = "lib"]
657 // Turn on a warning.
658 // This can be done in any module, not just the anonymous crate module.
659 #![warn(non_camel_case_types)]
662 A crate that contains a `main` function can be compiled to an executable. If a
663 `main` function is present, its return type must be [`unit`](#tuple-types)
664 and it must take no arguments.
666 # Items and attributes
668 Crates contain [items](#items), each of which may have some number of
669 [attributes](#attributes) attached to it.
673 An _item_ is a component of a crate. Items are organized within a crate by a
674 nested set of [modules](#modules). Every crate has a single "outermost"
675 anonymous module; all further items within the crate have [paths](#paths)
676 within the module tree of the crate.
678 Items are entirely determined at compile-time, generally remain fixed during
679 execution, and may reside in read-only memory.
681 There are several kinds of item:
683 * [`extern crate` declarations](#extern-crate-declarations)
684 * [`use` declarations](#use-declarations)
685 * [modules](#modules)
686 * [functions](#functions)
687 * [type definitions](grammar.html#type-definitions)
688 * [structs](#structs)
689 * [enumerations](#enumerations)
690 * [constant items](#constant-items)
691 * [static items](#static-items)
693 * [implementations](#implementations)
695 Some items form an implicit scope for the declaration of sub-items. In other
696 words, within a function or module, declarations of items can (in many cases)
697 be mixed with the statements, control blocks, and similar artifacts that
698 otherwise compose the item body. The meaning of these scoped items is the same
699 as if the item was declared outside the scope — it is still a static item
700 — except that the item's *path name* within the module namespace is
701 qualified by the name of the enclosing item, or is private to the enclosing
702 item (in the case of functions). The grammar specifies the exact locations in
703 which sub-item declarations may appear.
707 All items except modules, constants and statics may be *parameterized* by type.
708 Type parameters are given as a comma-separated list of identifiers enclosed in
709 angle brackets (`<...>`), after the name of the item and before its definition.
710 The type parameters of an item are considered "part of the name", not part of
711 the type of the item. A referencing [path](#paths) must (in principle) provide
712 type arguments as a list of comma-separated types enclosed within angle
713 brackets, in order to refer to the type-parameterized item. In practice, the
714 type-inference system can usually infer such argument types from context. There
715 are no general type-parametric types, only type-parametric items. That is, Rust
716 has no notion of type abstraction: there are no higher-ranked (or "forall") types
717 abstracted over other types, though higher-ranked types do exist for lifetimes.
721 A module is a container for zero or more [items](#items).
723 A _module item_ is a module, surrounded in braces, named, and prefixed with the
724 keyword `mod`. A module item introduces a new, named module into the tree of
725 modules making up a crate. Modules can nest arbitrarily.
727 An example of a module:
731 type Complex = (f64, f64);
732 fn sin(f: f64) -> f64 {
736 fn cos(f: f64) -> f64 {
740 fn tan(f: f64) -> f64 {
747 Modules and types share the same namespace. Declaring a named type with
748 the same name as a module in scope is forbidden: that is, a type definition,
749 trait, struct, enumeration, or type parameter can't shadow the name of a module
750 in scope, or vice versa.
752 A module without a body is loaded from an external file, by default with the
753 same name as the module, plus the `.rs` extension. When a nested submodule is
754 loaded from an external file, it is loaded from a subdirectory path that
755 mirrors the module hierarchy.
758 // Load the `vec` module from `vec.rs`
762 // Load the `local_data` module from `thread/local_data.rs`
763 // or `thread/local_data/mod.rs`.
768 The directories and files used for loading external file modules can be
769 influenced with the `path` attribute.
772 #[path = "thread_files"]
774 // Load the `local_data` module from `thread_files/tls.rs`
780 #### Extern crate declarations
782 An _`extern crate` declaration_ specifies a dependency on an external crate.
783 The external crate is then bound into the declaring scope as the `ident`
784 provided in the `extern_crate_decl`.
786 The external crate is resolved to a specific `soname` at compile time, and a
787 runtime linkage requirement to that `soname` is passed to the linker for
788 loading at runtime. The `soname` is resolved at compile time by scanning the
789 compiler's library path and matching the optional `crateid` provided against
790 the `crateid` attributes that were declared on the external crate when it was
791 compiled. If no `crateid` is provided, a default `name` attribute is assumed,
792 equal to the `ident` given in the `extern_crate_decl`.
794 Three examples of `extern crate` declarations:
799 extern crate std; // equivalent to: extern crate std as std;
801 extern crate std as ruststd; // linking to 'std' under another name
804 #### Use declarations
806 A _use declaration_ creates one or more local name bindings synonymous with
807 some other [path](#paths). Usually a `use` declaration is used to shorten the
808 path required to refer to a module item. These declarations may appear at the
809 top of [modules](#modules) and [blocks](grammar.html#block-expressions).
811 > **Note**: Unlike in many languages,
812 > `use` declarations in Rust do *not* declare linkage dependency with external crates.
813 > Rather, [`extern crate` declarations](#extern-crate-declarations) declare linkage dependencies.
815 Use declarations support a number of convenient shortcuts:
817 * Rebinding the target name as a new local name, using the syntax `use p::q::r as x;`
818 * Simultaneously binding a list of paths differing only in their final element,
819 using the glob-like brace syntax `use a::b::{c,d,e,f};`
820 * Binding all paths matching a given prefix, using the asterisk wildcard syntax
822 * Simultaneously binding a list of paths differing only in their final element
823 and their immediate parent module, using the `self` keyword, such as
824 `use a::b::{self, c, d};`
826 An example of `use` declarations:
829 use std::option::Option::{Some, None};
830 use std::collections::hash_map::{self, HashMap};
833 fn bar(map1: HashMap<String, usize>, map2: hash_map::HashMap<String, usize>){}
836 // Equivalent to 'foo(vec![std::option::Option::Some(1.0f64),
837 // std::option::Option::None]);'
838 foo(vec![Some(1.0f64), None]);
840 // Both `hash_map` and `HashMap` are in scope.
841 let map1 = HashMap::new();
842 let map2 = hash_map::HashMap::new();
847 Like items, `use` declarations are private to the containing module, by
848 default. Also like items, a `use` declaration can be public, if qualified by
849 the `pub` keyword. Such a `use` declaration serves to _re-export_ a name. A
850 public `use` declaration can therefore _redirect_ some public name to a
851 different target definition: even a definition with a private canonical path,
852 inside a different module. If a sequence of such redirections form a cycle or
853 cannot be resolved unambiguously, they represent a compile-time error.
855 An example of re-exporting:
860 pub use quux::foo::{bar, baz};
869 In this example, the module `quux` re-exports two public names defined in
872 Also note that the paths contained in `use` items are relative to the crate
873 root. So, in the previous example, the `use` refers to `quux::foo::{bar,
874 baz}`, and not simply to `foo::{bar, baz}`. This also means that top-level
875 module declarations should be at the crate root if direct usage of the declared
876 modules within `use` items is desired. It is also possible to use `self` and
877 `super` at the beginning of a `use` item to refer to the current and direct
878 parent modules respectively. All rules regarding accessing declared modules in
879 `use` declarations apply to both module declarations and `extern crate`
882 An example of what will and will not work for `use` items:
885 # #![allow(unused_imports)]
886 use foo::baz::foobaz; // good: foo is at the root of the crate
894 use foo::example::iter; // good: foo is at crate root
895 // use example::iter; // bad: example is not at the crate root
896 use self::baz::foobaz; // good: self refers to module 'foo'
897 use foo::bar::foobar; // good: foo is at crate root
904 use super::bar::foobar; // good: super refers to module 'foo'
914 A _function item_ defines a sequence of [statements](#statements) and a
915 final [expression](#expressions), along with a name and a set of
916 parameters. Other than a name, all these are optional.
917 Functions are declared with the keyword `fn`. Functions may declare a
918 set of *input* [*variables*](#variables) as parameters, through which the caller
919 passes arguments into the function, and the *output* [*type*](#types)
920 of the value the function will return to its caller on completion.
922 A function may also be copied into a first-class *value*, in which case the
923 value has the corresponding [*function type*](#function-types), and can be used
924 otherwise exactly as a function item (with a minor additional cost of calling
925 the function indirectly).
927 Every control path in a function logically ends with a `return` expression or a
928 diverging expression. If the outermost block of a function has a
929 value-producing expression in its final-expression position, that expression is
930 interpreted as an implicit `return` expression applied to the final-expression.
932 An example of a function:
935 fn add(x: i32, y: i32) -> i32 {
940 As with `let` bindings, function arguments are irrefutable patterns, so any
941 pattern that is valid in a let binding is also valid as an argument.
944 fn first((value, _): (i32, i32)) -> i32 { value }
948 #### Generic functions
950 A _generic function_ allows one or more _parameterized types_ to appear in its
951 signature. Each type parameter must be explicitly declared, in an
952 angle-bracket-enclosed, comma-separated list following the function name.
955 // foo is generic over A and B
957 fn foo<A, B>(x: A, y: B) {
960 Inside the function signature and body, the name of the type parameter can be
961 used as a type name. [Trait](#traits) bounds can be specified for type parameters
962 to allow methods with that trait to be called on values of that type. This is
963 specified using the `where` syntax:
966 fn foo<T>(x: T) where T: Debug {
969 When a generic function is referenced, its type is instantiated based on the
970 context of the reference. For example, calling the `foo` function here:
975 fn foo<T>(x: &[T]) where T: Debug {
983 will instantiate type parameter `T` with `i32`.
985 The type parameters can also be explicitly supplied in a trailing
986 [path](#paths) component after the function name. This might be necessary if
987 there is not sufficient context to determine the type parameters. For example,
988 `mem::size_of::<u32>() == 4`.
990 #### Diverging functions
992 A special kind of function can be declared with a `!` character where the
993 output type would normally be. For example:
996 fn my_err(s: &str) -> ! {
1002 We call such functions "diverging" because they never return a value to the
1003 caller. Every control path in a diverging function must end with a `panic!()` or
1004 a call to another diverging function on every control path. The `!` annotation
1005 does *not* denote a type.
1007 It might be necessary to declare a diverging function because as mentioned
1008 previously, the typechecker checks that every control path in a function ends
1009 with a [`return`](#return-expressions) or diverging expression. So, if `my_err`
1010 were declared without the `!` annotation, the following code would not
1014 # fn my_err(s: &str) -> ! { panic!() }
1016 fn f(i: i32) -> i32 {
1021 my_err("Bad number!");
1026 This will not compile without the `!` annotation on `my_err`, since the `else`
1027 branch of the conditional in `f` does not return an `i32`, as required by the
1028 signature of `f`. Adding the `!` annotation to `my_err` informs the
1029 typechecker that, should control ever enter `my_err`, no further type judgments
1030 about `f` need to hold, since control will never resume in any context that
1031 relies on those judgments. Thus the return type on `f` only needs to reflect
1032 the `if` branch of the conditional.
1034 #### Extern functions
1036 Extern functions are part of Rust's foreign function interface, providing the
1037 opposite functionality to [external blocks](#external-blocks). Whereas
1038 external blocks allow Rust code to call foreign code, extern functions with
1039 bodies defined in Rust code _can be called by foreign code_. They are defined
1040 in the same way as any other Rust function, except that they have the `extern`
1044 // Declares an extern fn, the ABI defaults to "C"
1045 extern fn new_i32() -> i32 { 0 }
1047 // Declares an extern fn with "stdcall" ABI
1048 extern "stdcall" fn new_i32_stdcall() -> i32 { 0 }
1051 Unlike normal functions, extern fns have type `extern "ABI" fn()`. This is the
1052 same type as the functions declared in an extern block.
1055 # extern fn new_i32() -> i32 { 0 }
1056 let fptr: extern "C" fn() -> i32 = new_i32;
1059 Extern functions may be called directly from Rust code as Rust uses large,
1060 contiguous stack segments like C.
1064 A _type alias_ defines a new name for an existing [type](#types). Type
1065 aliases are declared with the keyword `type`. Every value has a single,
1066 specific type, but may implement several different traits, or be compatible with
1067 several different type constraints.
1069 For example, the following defines the type `Point` as a synonym for the type
1070 `(u8, u8)`, the type of pairs of unsigned 8 bit integers:
1073 type Point = (u8, u8);
1074 let p: Point = (41, 68);
1079 A _struct_ is a nominal [struct type](#struct-types) defined with the
1082 An example of a `struct` item and its use:
1085 struct Point {x: i32, y: i32}
1086 let p = Point {x: 10, y: 11};
1090 A _tuple struct_ is a nominal [tuple type](#tuple-types), also defined with
1091 the keyword `struct`. For example:
1094 struct Point(i32, i32);
1095 let p = Point(10, 11);
1096 let px: i32 = match p { Point(x, _) => x };
1099 A _unit-like struct_ is a struct without any fields, defined by leaving off
1100 the list of fields entirely. Such a struct implicitly defines a constant of
1101 its type with the same name. For example:
1104 # #![feature(braced_empty_structs)]
1106 let c = [Cookie, Cookie {}, Cookie, Cookie {}];
1112 # #![feature(braced_empty_structs)]
1114 const Cookie: Cookie = Cookie {};
1115 let c = [Cookie, Cookie {}, Cookie, Cookie {}];
1118 The precise memory layout of a struct is not specified. One can specify a
1119 particular layout using the [`repr` attribute](#ffi-attributes).
1123 An _enumeration_ is a simultaneous definition of a nominal [enumerated
1124 type](#enumerated-types) as well as a set of *constructors*, that can be used
1125 to create or pattern-match values of the corresponding enumerated type.
1127 Enumerations are declared with the keyword `enum`.
1129 An example of an `enum` item and its use:
1137 let mut a: Animal = Animal::Dog;
1141 Enumeration constructors can have either named or unnamed fields:
1146 Cat { name: String, weight: f64 }
1149 let mut a: Animal = Animal::Dog("Cocoa".to_string(), 37.2);
1150 a = Animal::Cat { name: "Spotty".to_string(), weight: 2.7 };
1153 In this example, `Cat` is a _struct-like enum variant_,
1154 whereas `Dog` is simply called an enum variant.
1156 Enums have a discriminant. You can assign them explicitly:
1164 If a discriminant isn't assigned, they start at zero, and add one for each
1167 You can cast an enum to get this value:
1170 # enum Foo { Bar = 123 }
1171 let x = Foo::Bar as u32; // x is now 123u32
1174 This only works as long as none of the variants have data attached. If
1175 it were `Bar(i32)`, this is disallowed.
1179 A *constant item* is a named _constant value_ which is not associated with a
1180 specific memory location in the program. Constants are essentially inlined
1181 wherever they are used, meaning that they are copied directly into the relevant
1182 context when used. References to the same constant are not necessarily
1183 guaranteed to refer to the same memory address.
1185 Constant values must not have destructors, and otherwise permit most forms of
1186 data. Constants may refer to the address of other constants, in which case the
1187 address will have the `static` lifetime. The compiler is, however, still at
1188 liberty to translate the constant many times, so the address referred to may not
1191 Constants must be explicitly typed. The type may be `bool`, `char`, a number, or
1192 a type derived from those primitive types. The derived types are references with
1193 the `static` lifetime, fixed-size arrays, tuples, enum variants, and structs.
1196 const BIT1: u32 = 1 << 0;
1197 const BIT2: u32 = 1 << 1;
1199 const BITS: [u32; 2] = [BIT1, BIT2];
1200 const STRING: &'static str = "bitstring";
1202 struct BitsNStrings<'a> {
1207 const BITS_N_STRINGS: BitsNStrings<'static> = BitsNStrings {
1215 A *static item* is similar to a *constant*, except that it represents a precise
1216 memory location in the program. A static is never "inlined" at the usage site,
1217 and all references to it refer to the same memory location. Static items have
1218 the `static` lifetime, which outlives all other lifetimes in a Rust program.
1219 Static items may be placed in read-only memory if they do not contain any
1220 interior mutability.
1222 Statics may contain interior mutability through the `UnsafeCell` language item.
1223 All access to a static is safe, but there are a number of restrictions on
1226 * Statics may not contain any destructors.
1227 * The types of static values must ascribe to `Sync` to allow thread-safe access.
1228 * Statics may not refer to other statics by value, only by reference.
1229 * Constants cannot refer to statics.
1231 Constants should in general be preferred over statics, unless large amounts of
1232 data are being stored, or single-address and mutability properties are required.
1234 #### Mutable statics
1236 If a static item is declared with the `mut` keyword, then it is allowed to
1237 be modified by the program. One of Rust's goals is to make concurrency bugs
1238 hard to run into, and this is obviously a very large source of race conditions
1239 or other bugs. For this reason, an `unsafe` block is required when either
1240 reading or writing a mutable static variable. Care should be taken to ensure
1241 that modifications to a mutable static are safe with respect to other threads
1242 running in the same process.
1244 Mutable statics are still very useful, however. They can be used with C
1245 libraries and can also be bound from C libraries (in an `extern` block).
1248 # fn atomic_add(_: &mut u32, _: u32) -> u32 { 2 }
1250 static mut LEVELS: u32 = 0;
1252 // This violates the idea of no shared state, and this doesn't internally
1253 // protect against races, so this function is `unsafe`
1254 unsafe fn bump_levels_unsafe1() -> u32 {
1260 // Assuming that we have an atomic_add function which returns the old value,
1261 // this function is "safe" but the meaning of the return value may not be what
1262 // callers expect, so it's still marked as `unsafe`
1263 unsafe fn bump_levels_unsafe2() -> u32 {
1264 return atomic_add(&mut LEVELS, 1);
1268 Mutable statics have the same restrictions as normal statics, except that the
1269 type of the value is not required to ascribe to `Sync`.
1273 A _trait_ describes an abstract interface that types can
1274 implement. This interface consists of associated items, which come in
1281 Associated functions whose first parameter is named `self` are called
1282 methods and may be invoked using `.` notation (e.g., `x.foo()`).
1284 All traits define an implicit type parameter `Self` that refers to
1285 "the type that is implementing this interface". Traits may also
1286 contain additional type parameters. These type parameters (including
1287 `Self`) may be constrained by other traits and so forth as usual.
1289 Trait bounds on `Self` are considered "supertraits". These are
1290 required to be acyclic. Supertraits are somewhat different from other
1291 constraints in that they affect what methods are available in the
1292 vtable when the trait is used as a [trait object](#trait-objects).
1294 Traits are implemented for specific types through separate
1295 [implementations](#implementations).
1297 Consider the following trait:
1300 # type Surface = i32;
1301 # type BoundingBox = i32;
1303 fn draw(&self, Surface);
1304 fn bounding_box(&self) -> BoundingBox;
1308 This defines a trait with two methods. All values that have
1309 [implementations](#implementations) of this trait in scope can have their
1310 `draw` and `bounding_box` methods called, using `value.bounding_box()`
1311 [syntax](#method-call-expressions).
1313 Traits can include default implementations of methods, as in:
1318 fn baz(&self) { println!("We called baz."); }
1322 Here the `baz` method has a default implementation, so types that implement
1323 `Foo` need only implement `bar`. It is also possible for implementing types
1324 to override a method that has a default implementation.
1326 Type parameters can be specified for a trait to make it generic. These appear
1327 after the trait name, using the same syntax used in [generic
1328 functions](#generic-functions).
1332 fn len(&self) -> u32;
1333 fn elt_at(&self, n: u32) -> T;
1334 fn iter<F>(&self, F) where F: Fn(T);
1338 It is also possible to define associated types for a trait. Consider the
1339 following example of a `Container` trait. Notice how the type is available
1340 for use in the method signatures:
1346 fn insert(&mut self, Self::E);
1350 In order for a type to implement this trait, it must not only provide
1351 implementations for every method, but it must specify the type `E`. Here's
1352 an implementation of `Container` for the standard library type `Vec`:
1357 # fn empty() -> Self;
1358 # fn insert(&mut self, Self::E);
1360 impl<T> Container for Vec<T> {
1362 fn empty() -> Vec<T> { Vec::new() }
1363 fn insert(&mut self, x: T) { self.push(x); }
1367 Generic functions may use traits as _bounds_ on their type parameters. This
1368 will have two effects:
1370 - Only types that have the trait may instantiate the parameter.
1371 - Within the generic function, the methods of the trait can be
1372 called on values that have the parameter's type.
1377 # type Surface = i32;
1378 # trait Shape { fn draw(&self, Surface); }
1379 fn draw_twice<T: Shape>(surface: Surface, sh: T) {
1385 Traits also define a [trait object](#trait-objects) with the same
1386 name as the trait. Values of this type are created by coercing from a
1387 pointer of some specific type to a pointer of trait type. For example,
1388 `&T` could be coerced to `&Shape` if `T: Shape` holds (and similarly
1389 for `Box<T>`). This coercion can either be implicit or
1390 [explicit](#type-cast-expressions). Here is an example of an explicit
1395 impl Shape for i32 { }
1396 let mycircle = 0i32;
1397 let myshape: Box<Shape> = Box::new(mycircle) as Box<Shape>;
1400 The resulting value is a box containing the value that was cast, along with
1401 information that identifies the methods of the implementation that was used.
1402 Values with a trait type can have [methods called](#method-call-expressions) on
1403 them, for any method in the trait, and can be used to instantiate type
1404 parameters that are bounded by the trait.
1406 Trait methods may be static, which means that they lack a `self` argument.
1407 This means that they can only be called with function call syntax (`f(x)`) and
1408 not method call syntax (`obj.f()`). The way to refer to the name of a static
1409 method is to qualify it with the trait name, treating the trait name like a
1410 module. For example:
1414 fn from_i32(n: i32) -> Self;
1417 fn from_i32(n: i32) -> f64 { n as f64 }
1419 let x: f64 = Num::from_i32(42);
1422 Traits may inherit from other traits. Consider the following example:
1425 trait Shape { fn area(&self) -> f64; }
1426 trait Circle : Shape { fn radius(&self) -> f64; }
1429 The syntax `Circle : Shape` means that types that implement `Circle` must also
1430 have an implementation for `Shape`. Multiple supertraits are separated by `+`,
1431 `trait Circle : Shape + PartialEq { }`. In an implementation of `Circle` for a
1432 given type `T`, methods can refer to `Shape` methods, since the typechecker
1433 checks that any type with an implementation of `Circle` also has an
1434 implementation of `Shape`:
1439 trait Shape { fn area(&self) -> f64; }
1440 trait Circle : Shape { fn radius(&self) -> f64; }
1441 impl Shape for Foo {
1442 fn area(&self) -> f64 {
1446 impl Circle for Foo {
1447 fn radius(&self) -> f64 {
1448 println!("calling area: {}", self.area());
1458 In type-parameterized functions, methods of the supertrait may be called on
1459 values of subtrait-bound type parameters. Referring to the previous example of
1460 `trait Circle : Shape`:
1463 # trait Shape { fn area(&self) -> f64; }
1464 # trait Circle : Shape { fn radius(&self) -> f64; }
1465 fn radius_times_area<T: Circle>(c: T) -> f64 {
1466 // `c` is both a Circle and a Shape
1467 c.radius() * c.area()
1471 Likewise, supertrait methods may also be called on trait objects.
1474 # trait Shape { fn area(&self) -> f64; }
1475 # trait Circle : Shape { fn radius(&self) -> f64; }
1476 # impl Shape for i32 { fn area(&self) -> f64 { 0.0 } }
1477 # impl Circle for i32 { fn radius(&self) -> f64 { 0.0 } }
1478 # let mycircle = 0i32;
1479 let mycircle = Box::new(mycircle) as Box<Circle>;
1480 let nonsense = mycircle.radius() * mycircle.area();
1485 An _implementation_ is an item that implements a [trait](#traits) for a
1488 Implementations are defined with the keyword `impl`.
1491 # #[derive(Copy, Clone)]
1492 # struct Point {x: f64, y: f64};
1493 # type Surface = i32;
1494 # struct BoundingBox {x: f64, y: f64, width: f64, height: f64};
1495 # trait Shape { fn draw(&self, Surface); fn bounding_box(&self) -> BoundingBox; }
1496 # fn do_draw_circle(s: Surface, c: Circle) { }
1502 impl Copy for Circle {}
1504 impl Clone for Circle {
1505 fn clone(&self) -> Circle { *self }
1508 impl Shape for Circle {
1509 fn draw(&self, s: Surface) { do_draw_circle(s, *self); }
1510 fn bounding_box(&self) -> BoundingBox {
1511 let r = self.radius;
1513 x: self.center.x - r,
1514 y: self.center.y - r,
1522 It is possible to define an implementation without referring to a trait. The
1523 methods in such an implementation can only be used as direct calls on the values
1524 of the type that the implementation targets. In such an implementation, the
1525 trait type and `for` after `impl` are omitted. Such implementations are limited
1526 to nominal types (enums, structs, trait objects), and the implementation must
1527 appear in the same crate as the `self` type:
1530 struct Point {x: i32, y: i32}
1534 println!("Point is at ({}, {})", self.x, self.y);
1538 let my_point = Point {x: 10, y:11};
1542 When a trait _is_ specified in an `impl`, all methods declared as part of the
1543 trait must be implemented, with matching types and type parameter counts.
1545 An implementation can take type parameters, which can be different from the
1546 type parameters taken by the trait it implements. Implementation parameters
1547 are written after the `impl` keyword.
1550 # trait Seq<T> { fn dummy(&self, _: T) { } }
1551 impl<T> Seq<T> for Vec<T> {
1554 impl Seq<bool> for u32 {
1555 /* Treat the integer as a sequence of bits */
1561 External blocks form the basis for Rust's foreign function interface.
1562 Declarations in an external block describe symbols in external, non-Rust
1565 Functions within external blocks are declared in the same way as other Rust
1566 functions, with the exception that they may not have a body and are instead
1567 terminated by a semicolon.
1569 Functions within external blocks may be called by Rust code, just like
1570 functions defined in Rust. The Rust compiler automatically translates between
1571 the Rust ABI and the foreign ABI.
1573 A number of [attributes](#attributes) control the behavior of external blocks.
1575 By default external blocks assume that the library they are calling uses the
1576 standard C "cdecl" ABI. Other ABIs may be specified using an `abi` string, as
1580 // Interface to the Windows API
1581 extern "stdcall" { }
1584 The `link` attribute allows the name of the library to be specified. When
1585 specified the compiler will attempt to link against the native library of the
1589 #[link(name = "crypto")]
1593 The type of a function declared in an extern block is `extern "abi" fn(A1, ...,
1594 An) -> R`, where `A1...An` are the declared types of its arguments and `R` is
1595 the declared return type.
1597 It is valid to add the `link` attribute on an empty extern block. You can use
1598 this to satisfy the linking requirements of extern blocks elsewhere in your code
1599 (including upstream crates) instead of adding the attribute to each extern block.
1601 ## Visibility and Privacy
1603 These two terms are often used interchangeably, and what they are attempting to
1604 convey is the answer to the question "Can this item be used at this location?"
1606 Rust's name resolution operates on a global hierarchy of namespaces. Each level
1607 in the hierarchy can be thought of as some item. The items are one of those
1608 mentioned above, but also include external crates. Declaring or defining a new
1609 module can be thought of as inserting a new tree into the hierarchy at the
1610 location of the definition.
1612 To control whether interfaces can be used across modules, Rust checks each use
1613 of an item to see whether it should be allowed or not. This is where privacy
1614 warnings are generated, or otherwise "you used a private item of another module
1615 and weren't allowed to."
1617 By default, everything in Rust is *private*, with one exception. Enum variants
1618 in a `pub` enum are also public by default. When an item is declared as `pub`,
1619 it can be thought of as being accessible to the outside world. For example:
1623 // Declare a private struct
1626 // Declare a public struct with a private field
1631 // Declare a public enum with two public variants
1633 PubliclyAccessibleState,
1634 PubliclyAccessibleState2,
1638 With the notion of an item being either public or private, Rust allows item
1639 accesses in two cases:
1641 1. If an item is public, then it can be used externally through any of its
1643 2. If an item is private, it may be accessed by the current module and its
1646 These two cases are surprisingly powerful for creating module hierarchies
1647 exposing public APIs while hiding internal implementation details. To help
1648 explain, here's a few use cases and what they would entail:
1650 * A library developer needs to expose functionality to crates which link
1651 against their library. As a consequence of the first case, this means that
1652 anything which is usable externally must be `pub` from the root down to the
1653 destination item. Any private item in the chain will disallow external
1656 * A crate needs a global available "helper module" to itself, but it doesn't
1657 want to expose the helper module as a public API. To accomplish this, the
1658 root of the crate's hierarchy would have a private module which then
1659 internally has a "public API". Because the entire crate is a descendant of
1660 the root, then the entire local crate can access this private module through
1663 * When writing unit tests for a module, it's often a common idiom to have an
1664 immediate child of the module to-be-tested named `mod test`. This module
1665 could access any items of the parent module through the second case, meaning
1666 that internal implementation details could also be seamlessly tested from the
1669 In the second case, it mentions that a private item "can be accessed" by the
1670 current module and its descendants, but the exact meaning of accessing an item
1671 depends on what the item is. Accessing a module, for example, would mean
1672 looking inside of it (to import more items). On the other hand, accessing a
1673 function would mean that it is invoked. Additionally, path expressions and
1674 import statements are considered to access an item in the sense that the
1675 import/expression is only valid if the destination is in the current visibility
1678 Here's an example of a program which exemplifies the three cases outlined
1682 // This module is private, meaning that no external crate can access this
1683 // module. Because it is private at the root of this current crate, however, any
1684 // module in the crate may access any publicly visible item in this module.
1685 mod crate_helper_module {
1687 // This function can be used by anything in the current crate
1688 pub fn crate_helper() {}
1690 // This function *cannot* be used by anything else in the crate. It is not
1691 // publicly visible outside of the `crate_helper_module`, so only this
1692 // current module and its descendants may access it.
1693 fn implementation_detail() {}
1696 // This function is "public to the root" meaning that it's available to external
1697 // crates linking against this one.
1698 pub fn public_api() {}
1700 // Similarly to 'public_api', this module is public so external crates may look
1703 use crate_helper_module;
1705 pub fn my_method() {
1706 // Any item in the local crate may invoke the helper module's public
1707 // interface through a combination of the two rules above.
1708 crate_helper_module::crate_helper();
1711 // This function is hidden to any module which is not a descendant of
1713 fn my_implementation() {}
1719 fn test_my_implementation() {
1720 // Because this module is a descendant of `submodule`, it's allowed
1721 // to access private items inside of `submodule` without a privacy
1723 super::my_implementation();
1731 For a rust program to pass the privacy checking pass, all paths must be valid
1732 accesses given the two rules above. This includes all use statements,
1733 expressions, types, etc.
1735 ### Re-exporting and Visibility
1737 Rust allows publicly re-exporting items through a `pub use` directive. Because
1738 this is a public directive, this allows the item to be used in the current
1739 module through the rules above. It essentially allows public access into the
1740 re-exported item. For example, this program is valid:
1743 pub use self::implementation::api;
1745 mod implementation {
1754 This means that any external crate referencing `implementation::api::f` would
1755 receive a privacy violation, while the path `api::f` would be allowed.
1757 When re-exporting a private item, it can be thought of as allowing the "privacy
1758 chain" being short-circuited through the reexport instead of passing through
1759 the namespace hierarchy as it normally would.
1763 Any item declaration may have an _attribute_ applied to it. Attributes in Rust
1764 are modeled on Attributes in ECMA-335, with the syntax coming from ECMA-334
1765 (C#). An attribute is a general, free-form metadatum that is interpreted
1766 according to name, convention, and language and compiler version. Attributes
1767 may appear as any of:
1769 * A single identifier, the attribute name
1770 * An identifier followed by the equals sign '=' and a literal, providing a
1772 * An identifier followed by a parenthesized list of sub-attribute arguments
1774 Attributes with a bang ("!") after the hash ("#") apply to the item that the
1775 attribute is declared within. Attributes that do not have a bang after the hash
1776 apply to the item that follows the attribute.
1778 An example of attributes:
1781 // General metadata applied to the enclosing module or crate.
1782 #![crate_type = "lib"]
1784 // A function marked as a unit test
1790 // A conditionally-compiled module
1791 #[cfg(target_os="linux")]
1796 // A lint attribute used to suppress a warning/error
1797 #[allow(non_camel_case_types)]
1801 > **Note:** At some point in the future, the compiler will distinguish between
1802 > language-reserved and user-available attributes. Until then, there is
1803 > effectively no difference between an attribute handled by a loadable syntax
1804 > extension and the compiler.
1806 ### Crate-only attributes
1808 - `crate_name` - specify the crate's crate name.
1809 - `crate_type` - see [linkage](#linkage).
1810 - `feature` - see [compiler features](#compiler-features).
1811 - `no_builtins` - disable optimizing certain code patterns to invocations of
1812 library functions that are assumed to exist
1813 - `no_main` - disable emitting the `main` symbol. Useful when some other
1814 object being linked to defines `main`.
1815 - `no_start` - disable linking to the `native` crate, which specifies the
1816 "start" language item.
1817 - `no_std` - disable linking to the `std` crate.
1818 - `plugin` - load a list of named crates as compiler plugins, e.g.
1819 `#![plugin(foo, bar)]`. Optional arguments for each plugin,
1820 i.e. `#![plugin(foo(... args ...))]`, are provided to the plugin's
1821 registrar function. The `plugin` feature gate is required to use
1823 - `recursion_limit` - Sets the maximum depth for potentially
1824 infinitely-recursive compile-time operations like
1825 auto-dereference or macro expansion. The default is
1826 `#![recursion_limit="64"]`.
1828 ### Module-only attributes
1830 - `no_implicit_prelude` - disable injecting `use std::prelude::*` in this
1832 - `path` - specifies the file to load the module from. `#[path="foo.rs"] mod
1833 bar;` is equivalent to `mod bar { /* contents of foo.rs */ }`. The path is
1834 taken relative to the directory that the current module is in.
1836 ### Function-only attributes
1838 - `main` - indicates that this function should be passed to the entry point,
1839 rather than the function in the crate root named `main`.
1840 - `plugin_registrar` - mark this function as the registration point for
1841 [compiler plugins][plugin], such as loadable syntax extensions.
1842 - `start` - indicates that this function should be used as the entry point,
1843 overriding the "start" language item. See the "start" [language
1844 item](#language-items) for more details.
1845 - `test` - indicates that this function is a test function, to only be compiled
1846 in case of `--test`.
1847 - `should_panic` - indicates that this test function should panic, inverting the success condition.
1848 - `cold` - The function is unlikely to be executed, so optimize it (and calls
1851 ### Static-only attributes
1853 - `thread_local` - on a `static mut`, this signals that the value of this
1854 static may change depending on the current thread. The exact consequences of
1855 this are implementation-defined.
1859 On an `extern` block, the following attributes are interpreted:
1861 - `link_args` - specify arguments to the linker, rather than just the library
1862 name and type. This is feature gated and the exact behavior is
1863 implementation-defined (due to variety of linker invocation syntax).
1864 - `link` - indicate that a native library should be linked to for the
1865 declarations in this block to be linked correctly. `link` supports an optional
1866 `kind` key with three possible values: `dylib`, `static`, and `framework`. See
1867 [external blocks](#external-blocks) for more about external blocks. Two
1868 examples: `#[link(name = "readline")]` and
1869 `#[link(name = "CoreFoundation", kind = "framework")]`.
1870 - `linked_from` - indicates what native library this block of FFI items is
1871 coming from. This attribute is of the form `#[linked_from = "foo"]` where
1872 `foo` is the name of a library in either `#[link]` or a `-l` flag. This
1873 attribute is currently required to export symbols from a Rust dynamic library
1874 on Windows, and it is feature gated behind the `linked_from` feature.
1876 On declarations inside an `extern` block, the following attributes are
1879 - `link_name` - the name of the symbol that this function or static should be
1881 - `linkage` - on a static, this specifies the [linkage
1882 type](http://llvm.org/docs/LangRef.html#linkage-types).
1886 - `repr` - on C-like enums, this sets the underlying type used for
1887 representation. Takes one argument, which is the primitive
1888 type this enum should be represented for, or `C`, which specifies that it
1889 should be the default `enum` size of the C ABI for that platform. Note that
1890 enum representation in C is undefined, and this may be incorrect when the C
1891 code is compiled with certain flags.
1895 - `repr` - specifies the representation to use for this struct. Takes a list
1896 of options. The currently accepted ones are `C` and `packed`, which may be
1897 combined. `C` will use a C ABI compatible struct layout, and `packed` will
1898 remove any padding between fields (note that this is very fragile and may
1899 break platforms which require aligned access).
1901 ### Macro-related attributes
1903 - `macro_use` on a `mod` — macros defined in this module will be visible in the
1904 module's parent, after this module has been included.
1906 - `macro_use` on an `extern crate` — load macros from this crate. An optional
1907 list of names `#[macro_use(foo, bar)]` restricts the import to just those
1908 macros named. The `extern crate` must appear at the crate root, not inside
1909 `mod`, which ensures proper function of the [`$crate` macro
1910 variable](book/macros.html#the-variable-crate).
1912 - `macro_reexport` on an `extern crate` — re-export the named macros.
1914 - `macro_export` - export a macro for cross-crate usage.
1916 - `no_link` on an `extern crate` — even if we load this crate for macros, don't
1917 link it into the output.
1919 See the [macros section of the
1920 book](book/macros.html#scoping-and-macro-importexport) for more information on
1924 ### Miscellaneous attributes
1926 - `export_name` - on statics and functions, this determines the name of the
1928 - `link_section` - on statics and functions, this specifies the section of the
1929 object file that this item's contents will be placed into.
1930 - `no_mangle` - on any item, do not apply the standard name mangling. Set the
1931 symbol for this item to its identifier.
1932 - `simd` - on certain tuple structs, derive the arithmetic operators, which
1933 lower to the target's SIMD instructions, if any; the `simd` feature gate
1934 is necessary to use this attribute.
1935 - `unsafe_destructor_blind_to_params` - on `Drop::drop` method, asserts that the
1936 destructor code (and all potential specializations of that code) will
1937 never attempt to read from nor write to any references with lifetimes
1938 that come in via generic parameters. This is a constraint we cannot
1939 currently express via the type system, and therefore we rely on the
1940 programmer to assert that it holds. Adding this to a Drop impl causes
1941 the associated destructor to be considered "uninteresting" by the
1942 Drop-Check rule, and thus it can help sidestep data ordering
1943 constraints that would otherwise be introduced by the Drop-Check
1944 rule. Such sidestepping of the constraints, if done incorrectly, can
1945 lead to undefined behavior (in the form of reading or writing to data
1946 outside of its dynamic extent), and thus this attribute has the word
1947 "unsafe" in its name. To use this, the
1948 `unsafe_destructor_blind_to_params` feature gate must be enabled.
1949 - `unsafe_no_drop_flag` - on structs, remove the flag that prevents
1950 destructors from being run twice. Destructors might be run multiple times on
1951 the same object with this attribute. To use this, the `unsafe_no_drop_flag` feature
1952 gate must be enabled.
1953 - `doc` - Doc comments such as `/// foo` are equivalent to `#[doc = "foo"]`.
1954 - `rustc_on_unimplemented` - Write a custom note to be shown along with the error
1955 when the trait is found to be unimplemented on a type.
1956 You may use format arguments like `{T}`, `{A}` to correspond to the
1957 types at the point of use corresponding to the type parameters of the
1958 trait of the same name. `{Self}` will be replaced with the type that is supposed
1959 to implement the trait but doesn't. To use this, the `on_unimplemented` feature gate
1962 ### Conditional compilation
1964 Sometimes one wants to have different compiler outputs from the same code,
1965 depending on build target, such as targeted operating system, or to enable
1968 There are two kinds of configuration options, one that is either defined or not
1969 (`#[cfg(foo)]`), and the other that contains a string that can be checked
1970 against (`#[cfg(bar = "baz")]`). Currently, only compiler-defined configuration
1971 options can have the latter form.
1974 // The function is only included in the build when compiling for OSX
1975 #[cfg(target_os = "macos")]
1980 // This function is only included when either foo or bar is defined
1981 #[cfg(any(foo, bar))]
1982 fn needs_foo_or_bar() {
1986 // This function is only included when compiling for a unixish OS with a 32-bit
1988 #[cfg(all(unix, target_pointer_width = "32"))]
1989 fn on_32bit_unix() {
1993 // This function is only included when foo is not defined
1995 fn needs_not_foo() {
2000 This illustrates some conditional compilation can be achieved using the
2001 `#[cfg(...)]` attribute. `any`, `all` and `not` can be used to assemble
2002 arbitrarily complex configurations through nesting.
2004 The following configurations must be defined by the implementation:
2006 * `debug_assertions` - Enabled by default when compiling without optimizations.
2007 This can be used to enable extra debugging code in development but not in
2008 production. For example, it controls the behavior of the standard library's
2009 `debug_assert!` macro.
2010 * `target_arch = "..."` - Target CPU architecture, such as `"x86"`, `"x86_64"`
2011 `"mips"`, `"powerpc"`, `"arm"`, or `"aarch64"`.
2012 * `target_endian = "..."` - Endianness of the target CPU, either `"little"` or
2014 * `target_env = ".."` - An option provided by the compiler by default
2015 describing the runtime environment of the target platform. Some examples of
2016 this are `musl` for builds targeting the MUSL libc implementation, `msvc` for
2017 Windows builds targeting MSVC, and `gnu` frequently the rest of the time. This
2018 option may also be blank on some platforms.
2019 * `target_family = "..."` - Operating system family of the target, e. g.
2020 `"unix"` or `"windows"`. The value of this configuration option is defined
2021 as a configuration itself, like `unix` or `windows`.
2022 * `target_os = "..."` - Operating system of the target, examples include
2023 `"windows"`, `"macos"`, `"ios"`, `"linux"`, `"android"`, `"freebsd"`, `"dragonfly"`,
2024 `"bitrig"` , `"openbsd"` or `"netbsd"`.
2025 * `target_pointer_width = "..."` - Target pointer width in bits. This is set
2026 to `"32"` for targets with 32-bit pointers, and likewise set to `"64"` for
2028 * `target_vendor = "..."` - Vendor of the target, for example `apple`, `pc`, or
2030 * `test` - Enabled when compiling the test harness (using the `--test` flag).
2031 * `unix` - See `target_family`.
2032 * `windows` - See `target_family`.
2034 You can also set another attribute based on a `cfg` variable with `cfg_attr`:
2040 Will be the same as `#[b]` if `a` is set by `cfg`, and nothing otherwise.
2042 ### Lint check attributes
2044 A lint check names a potentially undesirable coding pattern, such as
2045 unreachable code or omitted documentation, for the static entity to which the
2048 For any lint check `C`:
2050 * `allow(C)` overrides the check for `C` so that violations will go
2052 * `deny(C)` signals an error after encountering a violation of `C`,
2053 * `forbid(C)` is the same as `deny(C)`, but also forbids changing the lint
2055 * `warn(C)` warns about violations of `C` but continues compilation.
2057 The lint checks supported by the compiler can be found via `rustc -W help`,
2058 along with their default settings. [Compiler
2059 plugins](book/compiler-plugins.html#lint-plugins) can provide additional lint checks.
2063 // Missing documentation is ignored here
2064 #[allow(missing_docs)]
2065 pub fn undocumented_one() -> i32 { 1 }
2067 // Missing documentation signals a warning here
2068 #[warn(missing_docs)]
2069 pub fn undocumented_too() -> i32 { 2 }
2071 // Missing documentation signals an error here
2072 #[deny(missing_docs)]
2073 pub fn undocumented_end() -> i32 { 3 }
2077 This example shows how one can use `allow` and `warn` to toggle a particular
2081 #[warn(missing_docs)]
2083 #[allow(missing_docs)]
2085 // Missing documentation is ignored here
2086 pub fn undocumented_one() -> i32 { 1 }
2088 // Missing documentation signals a warning here,
2089 // despite the allow above.
2090 #[warn(missing_docs)]
2091 pub fn undocumented_two() -> i32 { 2 }
2094 // Missing documentation signals a warning here
2095 pub fn undocumented_too() -> i32 { 3 }
2099 This example shows how one can use `forbid` to disallow uses of `allow` for
2103 #[forbid(missing_docs)]
2105 // Attempting to toggle warning signals an error here
2106 #[allow(missing_docs)]
2108 pub fn undocumented_too() -> i32 { 2 }
2114 Some primitive Rust operations are defined in Rust code, rather than being
2115 implemented directly in C or assembly language. The definitions of these
2116 operations have to be easy for the compiler to find. The `lang` attribute
2117 makes it possible to declare these operations. For example, the `str` module
2118 in the Rust standard library defines the string equality function:
2122 pub fn eq_slice(a: &str, b: &str) -> bool {
2127 The name `str_eq` has a special meaning to the Rust compiler, and the presence
2128 of this definition means that it will use this definition when generating calls
2129 to the string equality function.
2131 The set of language items is currently considered unstable. A complete
2132 list of the built-in language items will be added in the future.
2134 ### Inline attributes
2136 The inline attribute suggests that the compiler should place a copy of
2137 the function or static in the caller, rather than generating code to
2138 call the function or access the static where it is defined.
2140 The compiler automatically inlines functions based on internal heuristics.
2141 Incorrectly inlining functions can actually make the program slower, so it
2142 should be used with care.
2144 `#[inline]` and `#[inline(always)]` always cause the function to be serialized
2145 into the crate metadata to allow cross-crate inlining.
2147 There are three different types of inline attributes:
2149 * `#[inline]` hints the compiler to perform an inline expansion.
2150 * `#[inline(always)]` asks the compiler to always perform an inline expansion.
2151 * `#[inline(never)]` asks the compiler to never perform an inline expansion.
2155 The `derive` attribute allows certain traits to be automatically implemented
2156 for data structures. For example, the following will create an `impl` for the
2157 `PartialEq` and `Clone` traits for `Foo`, the type parameter `T` will be given
2158 the `PartialEq` or `Clone` constraints for the appropriate `impl`:
2161 #[derive(PartialEq, Clone)]
2168 The generated `impl` for `PartialEq` is equivalent to
2171 # struct Foo<T> { a: i32, b: T }
2172 impl<T: PartialEq> PartialEq for Foo<T> {
2173 fn eq(&self, other: &Foo<T>) -> bool {
2174 self.a == other.a && self.b == other.b
2177 fn ne(&self, other: &Foo<T>) -> bool {
2178 self.a != other.a || self.b != other.b
2183 ### Compiler Features
2185 Certain aspects of Rust may be implemented in the compiler, but they're not
2186 necessarily ready for every-day use. These features are often of "prototype
2187 quality" or "almost production ready", but may not be stable enough to be
2188 considered a full-fledged language feature.
2190 For this reason, Rust recognizes a special crate-level attribute of the form:
2193 #![feature(feature1, feature2, feature3)]
2196 This directive informs the compiler that the feature list: `feature1`,
2197 `feature2`, and `feature3` should all be enabled. This is only recognized at a
2198 crate-level, not at a module-level. Without this directive, all features are
2199 considered off, and using the features will result in a compiler error.
2201 The currently implemented features of the reference compiler are:
2203 * `advanced_slice_patterns` - See the [match expressions](#match-expressions)
2204 section for discussion; the exact semantics of
2205 slice patterns are subject to change, so some types
2208 * `slice_patterns` - OK, actually, slice patterns are just scary and
2209 completely unstable.
2211 * `asm` - The `asm!` macro provides a means for inline assembly. This is often
2212 useful, but the exact syntax for this feature along with its
2213 semantics are likely to change, so this macro usage must be opted
2216 * `associated_consts` - Allows constants to be defined in `impl` and `trait`
2217 blocks, so that they can be associated with a type or
2218 trait in a similar manner to methods and associated
2221 * `box_patterns` - Allows `box` patterns, the exact semantics of which
2222 is subject to change.
2224 * `box_syntax` - Allows use of `box` expressions, the exact semantics of which
2225 is subject to change.
2227 * `cfg_target_vendor` - Allows conditional compilation using the `target_vendor`
2228 matcher which is subject to change.
2230 * `concat_idents` - Allows use of the `concat_idents` macro, which is in many
2231 ways insufficient for concatenating identifiers, and may be
2232 removed entirely for something more wholesome.
2234 * `custom_attribute` - Allows the usage of attributes unknown to the compiler
2235 so that new attributes can be added in a backwards compatible
2238 * `custom_derive` - Allows the use of `#[derive(Foo,Bar)]` as sugar for
2239 `#[derive_Foo] #[derive_Bar]`, which can be user-defined syntax
2242 * `intrinsics` - Allows use of the "rust-intrinsics" ABI. Compiler intrinsics
2243 are inherently unstable and no promise about them is made.
2245 * `lang_items` - Allows use of the `#[lang]` attribute. Like `intrinsics`,
2246 lang items are inherently unstable and no promise about them
2249 * `link_args` - This attribute is used to specify custom flags to the linker,
2250 but usage is strongly discouraged. The compiler's usage of the
2251 system linker is not guaranteed to continue in the future, and
2252 if the system linker is not used then specifying custom flags
2253 doesn't have much meaning.
2255 * `link_llvm_intrinsics` – Allows linking to LLVM intrinsics via
2256 `#[link_name="llvm.*"]`.
2258 * `linkage` - Allows use of the `linkage` attribute, which is not portable.
2260 * `log_syntax` - Allows use of the `log_syntax` macro attribute, which is a
2261 nasty hack that will certainly be removed.
2263 * `main` - Allows use of the `#[main]` attribute, which changes the entry point
2264 into a Rust program. This capability is subject to change.
2266 * `macro_reexport` - Allows macros to be re-exported from one crate after being imported
2267 from another. This feature was originally designed with the sole
2268 use case of the Rust standard library in mind, and is subject to
2271 * `non_ascii_idents` - The compiler supports the use of non-ascii identifiers,
2272 but the implementation is a little rough around the
2273 edges, so this can be seen as an experimental feature
2274 for now until the specification of identifiers is fully
2277 * `no_std` - Allows the `#![no_std]` crate attribute, which disables the implicit
2278 `extern crate std`. This typically requires use of the unstable APIs
2279 behind the libstd "facade", such as libcore and libcollections. It
2280 may also cause problems when using syntax extensions, including
2283 * `on_unimplemented` - Allows the `#[rustc_on_unimplemented]` attribute, which allows
2284 trait definitions to add specialized notes to error messages
2285 when an implementation was expected but not found.
2287 * `optin_builtin_traits` - Allows the definition of default and negative trait
2288 implementations. Experimental.
2290 * `plugin` - Usage of [compiler plugins][plugin] for custom lints or syntax extensions.
2291 These depend on compiler internals and are subject to change.
2293 * `plugin_registrar` - Indicates that a crate provides [compiler plugins][plugin].
2295 * `quote` - Allows use of the `quote_*!` family of macros, which are
2296 implemented very poorly and will likely change significantly
2297 with a proper implementation.
2299 * `rustc_attrs` - Gates internal `#[rustc_*]` attributes which may be
2300 for internal use only or have meaning added to them in the future.
2302 * `rustc_diagnostic_macros`- A mysterious feature, used in the implementation
2303 of rustc, not meant for mortals.
2305 * `simd` - Allows use of the `#[simd]` attribute, which is overly simple and
2306 not the SIMD interface we want to expose in the long term.
2308 * `simd_ffi` - Allows use of SIMD vectors in signatures for foreign functions.
2309 The SIMD interface is subject to change.
2311 * `staged_api` - Allows usage of stability markers and `#![staged_api]` in a
2312 crate. Stability markers are also attributes: `#[stable]`,
2313 `#[unstable]`, and `#[deprecated]` are the three levels.
2315 * `start` - Allows use of the `#[start]` attribute, which changes the entry point
2316 into a Rust program. This capability, especially the signature for the
2317 annotated function, is subject to change.
2319 * `struct_variant` - Structural enum variants (those with named fields). It is
2320 currently unknown whether this style of enum variant is as
2321 fully supported as the tuple-forms, and it's not certain
2322 that this style of variant should remain in the language.
2323 For now this style of variant is hidden behind a feature
2326 * `thread_local` - The usage of the `#[thread_local]` attribute is experimental
2327 and should be seen as unstable. This attribute is used to
2328 declare a `static` as being unique per-thread leveraging
2329 LLVM's implementation which works in concert with the kernel
2330 loader and dynamic linker. This is not necessarily available
2331 on all platforms, and usage of it is discouraged.
2333 * `trace_macros` - Allows use of the `trace_macros` macro, which is a nasty
2334 hack that will certainly be removed.
2336 * `unboxed_closures` - Rust's new closure design, which is currently a work in
2337 progress feature with many known bugs.
2339 * `unsafe_no_drop_flag` - Allows use of the `#[unsafe_no_drop_flag]` attribute,
2340 which removes hidden flag added to a type that
2341 implements the `Drop` trait. The design for the
2342 `Drop` flag is subject to change, and this feature
2343 may be removed in the future.
2345 * `unmarked_api` - Allows use of items within a `#![staged_api]` crate
2346 which have not been marked with a stability marker.
2347 Such items should not be allowed by the compiler to exist,
2348 so if you need this there probably is a compiler bug.
2350 * `visible_private_types` - Allows public APIs to expose otherwise private
2351 types, e.g. as the return type of a public function.
2352 This capability may be removed in the future.
2354 * `allow_internal_unstable` - Allows `macro_rules!` macros to be tagged with the
2355 `#[allow_internal_unstable]` attribute, designed
2356 to allow `std` macros to call
2357 `#[unstable]`/feature-gated functionality
2358 internally without imposing on callers
2359 (i.e. making them behave like function calls in
2360 terms of encapsulation).
2361 * - `default_type_parameter_fallback` - Allows type parameter defaults to
2362 influence type inference.
2363 * - `braced_empty_structs` - Allows use of empty structs and enum variants with braces.
2365 If a feature is promoted to a language feature, then all existing programs will
2366 start to receive compilation warnings about `#![feature]` directives which enabled
2367 the new feature (because the directive is no longer necessary). However, if a
2368 feature is decided to be removed from the language, errors will be issued (if
2369 there isn't a parser error first). The directive in this case is no longer
2370 necessary, and it's likely that existing code will break if the feature isn't
2373 If an unknown feature is found in a directive, it results in a compiler error.
2374 An unknown feature is one which has never been recognized by the compiler.
2376 # Statements and expressions
2378 Rust is _primarily_ an expression language. This means that most forms of
2379 value-producing or effect-causing evaluation are directed by the uniform syntax
2380 category of _expressions_. Each kind of expression can typically _nest_ within
2381 each other kind of expression, and rules for evaluation of expressions involve
2382 specifying both the value produced by the expression and the order in which its
2383 sub-expressions are themselves evaluated.
2385 In contrast, statements in Rust serve _mostly_ to contain and explicitly
2386 sequence expression evaluation.
2390 A _statement_ is a component of a block, which is in turn a component of an
2391 outer [expression](#expressions) or [function](#functions).
2393 Rust has two kinds of statement: [declaration
2394 statements](#declaration-statements) and [expression
2395 statements](#expression-statements).
2397 ### Declaration statements
2399 A _declaration statement_ is one that introduces one or more *names* into the
2400 enclosing statement block. The declared names may denote new variables or new
2403 #### Item declarations
2405 An _item declaration statement_ has a syntactic form identical to an
2406 [item](#items) declaration within a module. Declaring an item — a
2407 function, enumeration, struct, type, static, trait, implementation or module
2408 — locally within a statement block is simply a way of restricting its
2409 scope to a narrow region containing all of its uses; it is otherwise identical
2410 in meaning to declaring the item outside the statement block.
2412 > **Note**: there is no implicit capture of the function's dynamic environment when
2413 > declaring a function-local item.
2415 #### Variable declarations
2417 A _variable declaration_ introduces a new set of variable, given by a pattern. The
2418 pattern may be followed by a type annotation, and/or an initializer expression.
2419 When no type annotation is given, the compiler will infer the type, or signal
2420 an error if insufficient type information is available for definite inference.
2421 Any variables introduced by a variable declaration are visible from the point of
2422 declaration until the end of the enclosing block scope.
2424 ### Expression statements
2426 An _expression statement_ is one that evaluates an [expression](#expressions)
2427 and ignores its result. The type of an expression statement `e;` is always
2428 `()`, regardless of the type of `e`. As a rule, an expression statement's
2429 purpose is to trigger the effects of evaluating its expression.
2433 An expression may have two roles: it always produces a *value*, and it may have
2434 *effects* (otherwise known as "side effects"). An expression *evaluates to* a
2435 value, and has effects during *evaluation*. Many expressions contain
2436 sub-expressions (operands). The meaning of each kind of expression dictates
2439 * Whether or not to evaluate the sub-expressions when evaluating the expression
2440 * The order in which to evaluate the sub-expressions
2441 * How to combine the sub-expressions' values to obtain the value of the expression
2443 In this way, the structure of expressions dictates the structure of execution.
2444 Blocks are just another kind of expression, so blocks, statements, expressions,
2445 and blocks again can recursively nest inside each other to an arbitrary depth.
2447 #### Lvalues, rvalues and temporaries
2449 Expressions are divided into two main categories: _lvalues_ and _rvalues_.
2450 Likewise within each expression, sub-expressions may occur in _lvalue context_
2451 or _rvalue context_. The evaluation of an expression depends both on its own
2452 category and the context it occurs within.
2454 An lvalue is an expression that represents a memory location. These expressions
2455 are [paths](#path-expressions) (which refer to local variables, function and
2456 method arguments, or static variables), dereferences (`*expr`), [indexing
2457 expressions](#index-expressions) (`expr[expr]`), and [field
2458 references](#field-expressions) (`expr.f`). All other expressions are rvalues.
2460 The left operand of an [assignment](#assignment-expressions) or
2461 [compound-assignment](#compound-assignment-expressions) expression is
2462 an lvalue context, as is the single operand of a unary
2463 [borrow](#unary-operator-expressions). The discriminant or subject of
2464 a [match expression](#match-expressions) may be an lvalue context, if
2465 ref bindings are made, but is otherwise an rvalue context. All other
2466 expression contexts are rvalue contexts.
2468 When an lvalue is evaluated in an _lvalue context_, it denotes a memory
2469 location; when evaluated in an _rvalue context_, it denotes the value held _in_
2470 that memory location.
2472 ##### Temporary lifetimes
2474 When an rvalue is used in an lvalue context, a temporary un-named
2475 lvalue is created and used instead. The lifetime of temporary values
2476 is typically the innermost enclosing statement; the tail expression of
2477 a block is considered part of the statement that encloses the block.
2479 When a temporary rvalue is being created that is assigned into a `let`
2480 declaration, however, the temporary is created with the lifetime of
2481 the enclosing block instead, as using the enclosing statement (the
2482 `let` declaration) would be a guaranteed error (since a pointer to the
2483 temporary would be stored into a variable, but the temporary would be
2484 freed before the variable could be used). The compiler uses simple
2485 syntactic rules to decide which values are being assigned into a `let`
2486 binding, and therefore deserve a longer temporary lifetime.
2488 Here are some examples:
2490 - `let x = foo(&temp())`. The expression `temp()` is an rvalue. As it
2491 is being borrowed, a temporary is created which will be freed after
2492 the innermost enclosing statement (the `let` declaration, in this case).
2493 - `let x = temp().foo()`. This is the same as the previous example,
2494 except that the value of `temp()` is being borrowed via autoref on a
2495 method-call. Here we are assuming that `foo()` is an `&self` method
2496 defined in some trait, say `Foo`. In other words, the expression
2497 `temp().foo()` is equivalent to `Foo::foo(&temp())`.
2498 - `let x = &temp()`. Here, the same temporary is being assigned into
2499 `x`, rather than being passed as a parameter, and hence the
2500 temporary's lifetime is considered to be the enclosing block.
2501 - `let x = SomeStruct { foo: &temp() }`. As in the previous case, the
2502 temporary is assigned into a struct which is then assigned into a
2503 binding, and hence it is given the lifetime of the enclosing block.
2504 - `let x = [ &temp() ]`. As in the previous case, the
2505 temporary is assigned into an array which is then assigned into a
2506 binding, and hence it is given the lifetime of the enclosing block.
2507 - `let ref x = temp()`. In this case, the temporary is created using a ref binding,
2508 but the result is the same: the lifetime is extended to the enclosing block.
2510 #### Moved and copied types
2512 When a [local variable](#variables) is used as an
2513 [rvalue](#lvalues-rvalues-and-temporaries), the variable will be copied
2514 if its type implements `Copy`. All others are moved.
2516 ### Literal expressions
2518 A _literal expression_ consists of one of the [literal](#literals) forms
2519 described earlier. It directly describes a number, character, string, boolean
2520 value, or the unit value.
2524 "hello"; // string type
2525 '5'; // character type
2529 ### Path expressions
2531 A [path](#paths) used as an expression context denotes either a local variable
2532 or an item. Path expressions are [lvalues](#lvalues-rvalues-and-temporaries).
2534 ### Tuple expressions
2536 Tuples are written by enclosing zero or more comma-separated expressions in
2537 parentheses. They are used to create [tuple-typed](#tuple-types) values.
2541 ("a", 4usize, true);
2544 You can disambiguate a single-element tuple from a value in parentheses with a
2548 (0,); // single-element tuple
2549 (0); // zero in parentheses
2552 ### Struct expressions
2554 There are several forms of struct expressions. A _struct expression_
2555 consists of the [path](#paths) of a [struct item](#structs), followed by
2556 a brace-enclosed list of one or more comma-separated name-value pairs,
2557 providing the field values of a new instance of the struct. A field name
2558 can be any identifier, and is separated from its value expression by a colon.
2559 The location denoted by a struct field is mutable if and only if the
2560 enclosing struct is mutable.
2562 A _tuple struct expression_ consists of the [path](#paths) of a [struct
2563 item](#structs), followed by a parenthesized list of one or more
2564 comma-separated expressions (in other words, the path of a struct item
2565 followed by a tuple expression). The struct item must be a tuple struct
2568 A _unit-like struct expression_ consists only of the [path](#paths) of a
2569 [struct item](#structs).
2571 The following are examples of struct expressions:
2574 # struct Point { x: f64, y: f64 }
2575 # struct TuplePoint(f64, f64);
2576 # mod game { pub struct User<'a> { pub name: &'a str, pub age: u32, pub score: usize } }
2577 # struct Cookie; fn some_fn<T>(t: T) {}
2578 Point {x: 10.0, y: 20.0};
2579 TuplePoint(10.0, 20.0);
2580 let u = game::User {name: "Joe", age: 35, score: 100_000};
2581 some_fn::<Cookie>(Cookie);
2584 A struct expression forms a new value of the named struct type. Note
2585 that for a given *unit-like* struct type, this will always be the same
2588 A struct expression can terminate with the syntax `..` followed by an
2589 expression to denote a functional update. The expression following `..` (the
2590 base) must have the same struct type as the new struct type being formed.
2591 The entire expression denotes the result of constructing a new struct (with
2592 the same type as the base expression) with the given values for the fields that
2593 were explicitly specified and the values in the base expression for all other
2597 # struct Point3d { x: i32, y: i32, z: i32 }
2598 let base = Point3d {x: 1, y: 2, z: 3};
2599 Point3d {y: 0, z: 10, .. base};
2602 ### Block expressions
2604 A _block expression_ is similar to a module in terms of the declarations that
2605 are possible. Each block conceptually introduces a new namespace scope. Use
2606 items can bring new names into scopes and declared items are in scope for only
2609 A block will execute each statement sequentially, and then execute the
2610 expression (if given). If the block ends in a statement, its value is `()`:
2613 let x: () = { println!("Hello."); };
2616 If it ends in an expression, its value and type are that of the expression:
2619 let x: i32 = { println!("Hello."); 5 };
2624 ### Method-call expressions
2626 A _method call_ consists of an expression followed by a single dot, an
2627 identifier, and a parenthesized expression-list. Method calls are resolved to
2628 methods on specific traits, either statically dispatching to a method if the
2629 exact `self`-type of the left-hand-side is known, or dynamically dispatching if
2630 the left-hand-side expression is an indirect [trait object](#trait-objects).
2632 ### Field expressions
2634 A _field expression_ consists of an expression followed by a single dot and an
2635 identifier, when not immediately followed by a parenthesized expression-list
2636 (the latter is a [method call expression](#method-call-expressions)). A field
2637 expression denotes a field of a [struct](#struct-types).
2642 (Struct {a: 10, b: 20}).a;
2645 A field access is an [lvalue](#lvalues-rvalues-and-temporaries) referring to
2646 the value of that field. When the type providing the field inherits mutability,
2647 it can be [assigned](#assignment-expressions) to.
2649 Also, if the type of the expression to the left of the dot is a
2650 pointer, it is automatically dereferenced as many times as necessary
2651 to make the field access possible. In cases of ambiguity, we prefer
2652 fewer autoderefs to more.
2654 ### Array expressions
2656 An [array](#array-and-slice-types) _expression_ is written by enclosing zero
2657 or more comma-separated expressions of uniform type in square brackets.
2659 In the `[expr ';' expr]` form, the expression after the `';'` must be a
2660 constant expression that can be evaluated at compile time, such as a
2661 [literal](#literals) or a [static item](#static-items).
2665 ["a", "b", "c", "d"];
2666 [0; 128]; // array with 128 zeros
2667 [0u8, 0u8, 0u8, 0u8];
2670 ### Index expressions
2672 [Array](#array-and-slice-types)-typed expressions can be indexed by
2673 writing a square-bracket-enclosed expression (the index) after them. When the
2674 array is mutable, the resulting [lvalue](#lvalues-rvalues-and-temporaries) can
2677 Indices are zero-based, and may be of any integral type. Vector access is
2678 bounds-checked at compile-time for constant arrays being accessed with a constant index value.
2679 Otherwise a check will be performed at run-time that will put the thread in a _panicked state_ if it fails.
2684 let x = (["a", "b"])[10]; // compiler error: const index-expr is out of bounds
2687 let y = (["a", "b"])[n]; // panics
2689 let arr = ["a", "b"];
2693 Also, if the type of the expression to the left of the brackets is a
2694 pointer, it is automatically dereferenced as many times as necessary
2695 to make the indexing possible. In cases of ambiguity, we prefer fewer
2698 ### Range expressions
2700 The `..` operator will construct an object of one of the `std::ops::Range` variants.
2703 1..2; // std::ops::Range
2704 3..; // std::ops::RangeFrom
2705 ..4; // std::ops::RangeTo
2706 ..; // std::ops::RangeFull
2709 The following expressions are equivalent.
2712 let x = std::ops::Range {start: 0, end: 10};
2718 ### Unary operator expressions
2720 Rust defines the following unary operators. They are all written as prefix operators,
2721 before the expression they apply to.
2724 : Negation. May only be applied to numeric types.
2726 : Dereference. When applied to a [pointer](#pointer-types) it denotes the
2727 pointed-to location. For pointers to mutable locations, the resulting
2728 [lvalue](#lvalues-rvalues-and-temporaries) can be assigned to.
2729 On non-pointer types, it calls the `deref` method of the `std::ops::Deref`
2730 trait, or the `deref_mut` method of the `std::ops::DerefMut` trait (if
2731 implemented by the type and required for an outer expression that will or
2732 could mutate the dereference), and produces the result of dereferencing the
2733 `&` or `&mut` borrowed pointer returned from the overload method.
2735 : Logical negation. On the boolean type, this flips between `true` and
2736 `false`. On integer types, this inverts the individual bits in the
2737 two's complement representation of the value.
2739 : Borrowing. When applied to an lvalue, these operators produce a
2740 reference (pointer) to the lvalue. The lvalue is also placed into
2741 a borrowed state for the duration of the reference. For a shared
2742 borrow (`&`), this implies that the lvalue may not be mutated, but
2743 it may be read or shared again. For a mutable borrow (`&mut`), the
2744 lvalue may not be accessed in any way until the borrow expires.
2745 If the `&` or `&mut` operators are applied to an rvalue, a
2746 temporary value is created; the lifetime of this temporary value
2747 is defined by [syntactic rules](#temporary-lifetimes).
2749 ### Binary operator expressions
2751 Binary operators expressions are given in terms of [operator
2752 precedence](#operator-precedence).
2754 #### Arithmetic operators
2756 Binary arithmetic expressions are syntactic sugar for calls to built-in traits,
2757 defined in the `std::ops` module of the `std` library. This means that
2758 arithmetic operators can be overridden for user-defined types. The default
2759 meaning of the operators on standard types is given here.
2762 : Addition and array/string concatenation.
2763 Calls the `add` method on the `std::ops::Add` trait.
2766 Calls the `sub` method on the `std::ops::Sub` trait.
2769 Calls the `mul` method on the `std::ops::Mul` trait.
2772 Calls the `div` method on the `std::ops::Div` trait.
2775 Calls the `rem` method on the `std::ops::Rem` trait.
2777 #### Bitwise operators
2779 Like the [arithmetic operators](#arithmetic-operators), bitwise operators are
2780 syntactic sugar for calls to methods of built-in traits. This means that
2781 bitwise operators can be overridden for user-defined types. The default
2782 meaning of the operators on standard types is given here. Bitwise `&`, `|` and
2783 `^` applied to boolean arguments are equivalent to logical `&&`, `||` and `!=`
2784 evaluated in non-lazy fashion.
2788 Calls the `bitand` method of the `std::ops::BitAnd` trait.
2790 : Bitwise inclusive OR.
2791 Calls the `bitor` method of the `std::ops::BitOr` trait.
2793 : Bitwise exclusive OR.
2794 Calls the `bitxor` method of the `std::ops::BitXor` trait.
2797 Calls the `shl` method of the `std::ops::Shl` trait.
2799 : Right shift (arithmetic).
2800 Calls the `shr` method of the `std::ops::Shr` trait.
2802 #### Lazy boolean operators
2804 The operators `||` and `&&` may be applied to operands of boolean type. The
2805 `||` operator denotes logical 'or', and the `&&` operator denotes logical
2806 'and'. They differ from `|` and `&` in that the right-hand operand is only
2807 evaluated when the left-hand operand does not already determine the result of
2808 the expression. That is, `||` only evaluates its right-hand operand when the
2809 left-hand operand evaluates to `false`, and `&&` only when it evaluates to
2812 #### Comparison operators
2814 Comparison operators are, like the [arithmetic
2815 operators](#arithmetic-operators), and [bitwise operators](#bitwise-operators),
2816 syntactic sugar for calls to built-in traits. This means that comparison
2817 operators can be overridden for user-defined types. The default meaning of the
2818 operators on standard types is given here.
2822 Calls the `eq` method on the `std::cmp::PartialEq` trait.
2825 Calls the `ne` method on the `std::cmp::PartialEq` trait.
2828 Calls the `lt` method on the `std::cmp::PartialOrd` trait.
2831 Calls the `gt` method on the `std::cmp::PartialOrd` trait.
2833 : Less than or equal.
2834 Calls the `le` method on the `std::cmp::PartialOrd` trait.
2836 : Greater than or equal.
2837 Calls the `ge` method on the `std::cmp::PartialOrd` trait.
2839 #### Type cast expressions
2841 A type cast expression is denoted with the binary operator `as`.
2843 Executing an `as` expression casts the value on the left-hand side to the type
2844 on the right-hand side.
2846 An example of an `as` expression:
2849 # fn sum(values: &[f64]) -> f64 { 0.0 }
2850 # fn len(values: &[f64]) -> i32 { 0 }
2852 fn average(values: &[f64]) -> f64 {
2853 let sum: f64 = sum(values);
2854 let size: f64 = len(values) as f64;
2859 Some of the conversions which can be done through the `as` operator
2860 can also be done implicitly at various points in the program, such as
2861 argument passing and assignment to a `let` binding with an explicit
2862 type. Implicit conversions are limited to "harmless" conversions that
2863 do not lose information and which have minimal or no risk of
2864 surprising side-effects on the dynamic execution semantics.
2866 #### Assignment expressions
2868 An _assignment expression_ consists of an
2869 [lvalue](#lvalues-rvalues-and-temporaries) expression followed by an equals
2870 sign (`=`) and an [rvalue](#lvalues-rvalues-and-temporaries) expression.
2872 Evaluating an assignment expression [either copies or
2873 moves](#moved-and-copied-types) its right-hand operand to its left-hand
2882 #### Compound assignment expressions
2884 The `+`, `-`, `*`, `/`, `%`, `&`, `|`, `^`, `<<`, and `>>` operators may be
2885 composed with the `=` operator. The expression `lval OP= val` is equivalent to
2886 `lval = lval OP val`. For example, `x = x + 1` may be written as `x += 1`.
2888 Any such expression always has the [`unit`](#tuple-types) type.
2890 #### Operator precedence
2892 The precedence of Rust binary operators is ordered as follows, going from
2895 ```{.text .precedence}
2909 Operators at the same precedence level are evaluated left-to-right. [Unary
2910 operators](#unary-operator-expressions) have the same precedence level and are
2911 stronger than any of the binary operators.
2913 ### Grouped expressions
2915 An expression enclosed in parentheses evaluates to the result of the enclosed
2916 expression. Parentheses can be used to explicitly specify evaluation order
2917 within an expression.
2919 An example of a parenthesized expression:
2922 let x: i32 = (2 + 3) * 4;
2926 ### Call expressions
2928 A _call expression_ invokes a function, providing zero or more input variables
2929 and an optional location to move the function's output into. If the function
2930 eventually returns, then the expression completes.
2932 Some examples of call expressions:
2935 # fn add(x: i32, y: i32) -> i32 { 0 }
2937 let x: i32 = add(1i32, 2i32);
2938 let pi: Result<f32, _> = "3.14".parse();
2941 ### Lambda expressions
2943 A _lambda expression_ (sometimes called an "anonymous function expression")
2944 defines a function and denotes it as a value, in a single expression. A lambda
2945 expression is a pipe-symbol-delimited (`|`) list of identifiers followed by an
2948 A lambda expression denotes a function that maps a list of parameters
2949 (`ident_list`) onto the expression that follows the `ident_list`. The
2950 identifiers in the `ident_list` are the parameters to the function. These
2951 parameters' types need not be specified, as the compiler infers them from
2954 Lambda expressions are most useful when passing functions as arguments to other
2955 functions, as an abbreviation for defining and capturing a separate function.
2957 Significantly, lambda expressions _capture their environment_, which regular
2958 [function definitions](#functions) do not. The exact type of capture depends
2959 on the [function type](#function-types) inferred for the lambda expression. In
2960 the simplest and least-expensive form (analogous to a ```|| { }``` expression),
2961 the lambda expression captures its environment by reference, effectively
2962 borrowing pointers to all outer variables mentioned inside the function.
2963 Alternately, the compiler may infer that a lambda expression should copy or
2964 move values (depending on their type) from the environment into the lambda
2965 expression's captured environment.
2967 In this example, we define a function `ten_times` that takes a higher-order
2968 function argument, and we then call it with a lambda expression as an argument:
2971 fn ten_times<F>(f: F) where F: Fn(i32) {
2972 for index in 0..10 {
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.
3035 ### `for` expressions
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.
3070 ### `if` expressions
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, structs 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 Multiple match patterns may be joined with the `|` operator. A range of values
3155 may be specified with `...`. For example:
3160 let message = match x {
3161 0 | 1 => "not many",
3167 Range patterns only work on scalar types (like integers and characters; not
3168 like arrays and structs, which have sub-components). A range pattern may not
3169 be a sub-range of another range pattern inside the same `match`.
3171 Finally, match patterns can accept *pattern guards* to further refine the
3172 criteria for matching a case. Pattern guards appear after the pattern and
3173 consist of a bool-typed expression following the `if` keyword. A pattern guard
3174 may refer to the variables bound within the pattern they follow.
3177 # let maybe_digit = Some(0);
3178 # fn process_digit(i: i32) { }
3179 # fn process_other(i: i32) { }
3181 let message = match maybe_digit {
3182 Some(x) if x < 10 => process_digit(x),
3183 Some(x) => process_other(x),
3188 ### `if let` expressions
3190 An `if let` expression is semantically identical to an `if` expression but in place
3191 of a condition expression it expects a refutable let statement. If the value of the
3192 expression on the right hand side of the let statement matches the pattern, the corresponding
3193 block will execute, otherwise flow proceeds to the first `else` block that follows.
3196 let dish = ("Ham", "Eggs");
3198 // this body will be skipped because the pattern is refuted
3199 if let ("Bacon", b) = dish {
3200 println!("Bacon is served with {}", b);
3203 // this body will execute
3204 if let ("Ham", b) = dish {
3205 println!("Ham is served with {}", b);
3209 ### `while let` loops
3211 A `while let` loop is semantically identical to a `while` loop but in place of a
3212 condition expression it expects a refutable let statement. If the value of the
3213 expression on the right hand side of the let statement matches the pattern, the
3214 loop body block executes and control returns to the pattern matching statement.
3215 Otherwise, the while expression completes.
3217 ### `return` expressions
3219 Return expressions are denoted with the keyword `return`. Evaluating a `return`
3220 expression moves its argument into the designated output location for the
3221 current function call, destroys the current function activation frame, and
3222 transfers control to the caller frame.
3224 An example of a `return` expression:
3227 fn max(a: i32, b: i32) -> i32 {
3239 Every variable, item and value in a Rust program has a type. The _type_ of a
3240 *value* defines the interpretation of the memory holding it.
3242 Built-in types and type-constructors are tightly integrated into the language,
3243 in nontrivial ways that are not possible to emulate in user-defined types.
3244 User-defined types have limited capabilities.
3248 The primitive types are the following:
3250 * The boolean type `bool` with values `true` and `false`.
3251 * The machine types (integer and floating-point).
3252 * The machine-dependent integer types.
3256 The machine types are the following:
3258 * The unsigned word types `u8`, `u16`, `u32` and `u64`, with values drawn from
3259 the integer intervals [0, 2^8 - 1], [0, 2^16 - 1], [0, 2^32 - 1] and
3260 [0, 2^64 - 1] respectively.
3262 * The signed two's complement word types `i8`, `i16`, `i32` and `i64`, with
3263 values drawn from the integer intervals [-(2^(7)), 2^7 - 1],
3264 [-(2^(15)), 2^15 - 1], [-(2^(31)), 2^31 - 1], [-(2^(63)), 2^63 - 1]
3267 * The IEEE 754-2008 `binary32` and `binary64` floating-point types: `f32` and
3268 `f64`, respectively.
3270 #### Machine-dependent integer types
3272 The `usize` type is an unsigned integer type with the same number of bits as the
3273 platform's pointer type. It can represent every memory address in the process.
3275 The `isize` type is a signed integer type with the same number of bits as the
3276 platform's pointer type. The theoretical upper bound on object and array size
3277 is the maximum `isize` value. This ensures that `isize` can be used to calculate
3278 differences between pointers into an object or array and can address every byte
3279 within an object along with one byte past the end.
3283 The types `char` and `str` hold textual data.
3285 A value of type `char` is a [Unicode scalar value](
3286 http://www.unicode.org/glossary/#unicode_scalar_value) (i.e. a code point that
3287 is not a surrogate), represented as a 32-bit unsigned word in the 0x0000 to
3288 0xD7FF or 0xE000 to 0x10FFFF range. A `[char]` array is effectively an UCS-4 /
3291 A value of type `str` is a Unicode string, represented as an array of 8-bit
3292 unsigned bytes holding a sequence of UTF-8 code points. Since `str` is of
3293 unknown size, it is not a _first-class_ type, but can only be instantiated
3294 through a pointer type, such as `&str`.
3298 A tuple *type* is a heterogeneous product of other types, called the *elements*
3299 of the tuple. It has no nominal name and is instead structurally typed.
3301 Tuple types and values are denoted by listing the types or values of their
3302 elements, respectively, in a parenthesized, comma-separated list.
3304 Because tuple elements don't have a name, they can only be accessed by
3305 pattern-matching or by using `N` directly as a field to access the
3308 An example of a tuple type and its use:
3311 type Pair<'a> = (i32, &'a str);
3312 let p: Pair<'static> = (10, "ten");
3316 assert_eq!(b, "ten");
3317 assert_eq!(p.0, 10);
3318 assert_eq!(p.1, "ten");
3321 For historical reasons and convenience, the tuple type with no elements (`()`)
3322 is often called ‘unit’ or ‘the unit type’.
3324 ### Array, and Slice types
3326 Rust has two different types for a list of items:
3328 * `[T; N]`, an 'array'
3331 An array has a fixed size, and can be allocated on either the stack or the
3334 A slice is a 'view' into an array. It doesn't own the data it points
3340 // A stack-allocated array
3341 let array: [i32; 3] = [1, 2, 3];
3343 // A heap-allocated array
3344 let vector: Vec<i32> = vec![1, 2, 3];
3346 // A slice into an array
3347 let slice: &[i32] = &vector[..];
3350 As you can see, the `vec!` macro allows you to create a `Vec<T>` easily. The
3351 `vec!` macro is also part of the standard library, rather than the language.
3353 All in-bounds elements of arrays and slices are always initialized, and access
3354 to an array or slice is always bounds-checked.
3358 A `struct` *type* is a heterogeneous product of other types, called the
3359 *fields* of the type.[^structtype]
3361 [^structtype]: `struct` types are analogous to `struct` types in C,
3362 the *record* types of the ML family,
3363 or the *struct* types of the Lisp family.
3365 New instances of a `struct` can be constructed with a [struct
3366 expression](#struct-expressions).
3368 The memory layout of a `struct` is undefined by default to allow for compiler
3369 optimizations like field reordering, but it can be fixed with the
3370 `#[repr(...)]` attribute. In either case, fields may be given in any order in
3371 a corresponding struct *expression*; the resulting `struct` value will always
3372 have the same memory layout.
3374 The fields of a `struct` may be qualified by [visibility
3375 modifiers](#visibility-and-privacy), to allow access to data in a
3376 struct outside a module.
3378 A _tuple struct_ type is just like a struct type, except that the fields are
3381 A _unit-like struct_ type is like a struct type, except that it has no
3382 fields. The one value constructed by the associated [struct
3383 expression](#struct-expressions) is the only value that inhabits such a
3386 ### Enumerated types
3388 An *enumerated type* is a nominal, heterogeneous disjoint union type, denoted
3389 by the name of an [`enum` item](#enumerations). [^enumtype]
3391 [^enumtype]: The `enum` type is analogous to a `data` constructor declaration in
3392 ML, or a *pick ADT* in Limbo.
3394 An [`enum` item](#enumerations) declares both the type and a number of *variant
3395 constructors*, each of which is independently named and takes an optional tuple
3398 New instances of an `enum` can be constructed by calling one of the variant
3399 constructors, in a [call expression](#call-expressions).
3401 Any `enum` value consumes as much memory as the largest variant constructor for
3402 its corresponding `enum` type.
3404 Enum types cannot be denoted *structurally* as types, but must be denoted by
3405 named reference to an [`enum` item](#enumerations).
3409 Nominal types — [enumerations](#enumerated-types) and
3410 [structs](#struct-types) — may be recursive. That is, each `enum`
3411 constructor or `struct` field may refer, directly or indirectly, to the
3412 enclosing `enum` or `struct` type itself. Such recursion has restrictions:
3414 * Recursive types must include a nominal type in the recursion
3415 (not mere [type definitions](grammar.html#type-definitions),
3416 or other structural types such as [arrays](#array-and-slice-types) or [tuples](#tuple-types)).
3417 * A recursive `enum` item must have at least one non-recursive constructor
3418 (in order to give the recursion a basis case).
3419 * The size of a recursive type must be finite;
3420 in other words the recursive fields of the type must be [pointer types](#pointer-types).
3421 * Recursive type definitions can cross module boundaries, but not module *visibility* boundaries,
3422 or crate boundaries (in order to simplify the module system and type checker).
3424 An example of a *recursive* type and its use:
3429 Cons(T, Box<List<T>>)
3432 let a: List<i32> = List::Cons(7, Box::new(List::Cons(13, Box::new(List::Nil))));
3437 All pointers in Rust are explicit first-class values. They can be copied,
3438 stored into data structs, and returned from functions. There are two
3439 varieties of pointer in Rust:
3442 : These point to memory _owned by some other value_.
3443 A reference type is written `&type`,
3444 or `&'a type` when you need to specify an explicit lifetime.
3445 Copying a reference is a "shallow" operation:
3446 it involves only copying the pointer itself.
3447 Releasing a reference has no effect on the value it points to,
3448 but a reference of a temporary value will keep it alive during the scope
3449 of the reference itself.
3451 * Raw pointers (`*`)
3452 : Raw pointers are pointers without safety or liveness guarantees.
3453 Raw pointers are written as `*const T` or `*mut T`,
3454 for example `*const i32` means a raw pointer to a 32-bit integer.
3455 Copying or dropping a raw pointer has no effect on the lifecycle of any
3456 other value. Dereferencing a raw pointer or converting it to any other
3457 pointer type is an [`unsafe` operation](#unsafe-functions).
3458 Raw pointers are generally discouraged in Rust code;
3459 they exist to support interoperability with foreign code,
3460 and writing performance-critical or low-level functions.
3462 The standard library contains additional 'smart pointer' types beyond references
3467 The function type constructor `fn` forms new function types. A function type
3468 consists of a possibly-empty set of function-type modifiers (such as `unsafe`
3469 or `extern`), a sequence of input types and an output type.
3471 An example of a `fn` type:
3474 fn add(x: i32, y: i32) -> i32 {
3478 let mut x = add(5,7);
3480 type Binop = fn(i32, i32) -> i32;
3481 let bo: Binop = add;
3485 #### Function types for specific items
3487 Internal to the compiler, there are also function types that are specific to a particular
3488 function item. In the following snippet, for example, the internal types of the functions
3489 `foo` and `bar` are different, despite the fact that they have the same signature:
3496 The types of `foo` and `bar` can both be implicitly coerced to the fn
3497 pointer type `fn()`. There is currently no syntax for unique fn types,
3498 though the compiler will emit a type like `fn() {foo}` in error
3499 messages to indicate "the unique fn type for the function `foo`".
3503 A [lambda expression](#lambda-expressions) produces a closure value with
3504 a unique, anonymous type that cannot be written out.
3506 Depending on the requirements of the closure, its type implements one or
3507 more of the closure traits:
3510 : The closure can be called once. A closure called as `FnOnce`
3511 can move out values from its environment.
3514 : The closure can be called multiple times as mutable. A closure called as
3515 `FnMut` can mutate values from its environment. `FnMut` inherits from
3516 `FnOnce` (i.e. anything implementing `FnMut` also implements `FnOnce`).
3519 : The closure can be called multiple times through a shared reference.
3520 A closure called as `Fn` can neither move out from nor mutate values
3521 from its environment. `Fn` inherits from `FnMut`, which itself
3522 inherits from `FnOnce`.
3527 In Rust, a type like `&SomeTrait` or `Box<SomeTrait>` is called a _trait object_.
3528 Each instance of a trait object includes:
3530 - a pointer to an instance of a type `T` that implements `SomeTrait`
3531 - a _virtual method table_, often just called a _vtable_, which contains, for
3532 each method of `SomeTrait` that `T` implements, a pointer to `T`'s
3533 implementation (i.e. a function pointer).
3535 The purpose of trait objects is to permit "late binding" of methods. A call to
3536 a method on a trait object is only resolved to a vtable entry at compile time.
3537 The actual implementation for each vtable entry can vary on an object-by-object
3540 Note that for a trait object to be instantiated, the trait must be
3541 _object-safe_. Object safety rules are defined in [RFC 255].
3543 [RFC 255]: https://github.com/rust-lang/rfcs/blob/master/text/0255-object-safety.md
3545 Given a pointer-typed expression `E` of type `&T` or `Box<T>`, where `T`
3546 implements trait `R`, casting `E` to the corresponding pointer type `&R` or
3547 `Box<R>` results in a value of the _trait object_ `R`. This result is
3548 represented as a pair of pointers: the vtable pointer for the `T`
3549 implementation of `R`, and the pointer value of `E`.
3551 An example of a trait object:
3555 fn stringify(&self) -> String;
3558 impl Printable for i32 {
3559 fn stringify(&self) -> String { self.to_string() }
3562 fn print(a: Box<Printable>) {
3563 println!("{}", a.stringify());
3567 print(Box::new(10) as Box<Printable>);
3571 In this example, the trait `Printable` occurs as a trait object in both the
3572 type signature of `print`, and the cast expression in `main`.
3576 Within the body of an item that has type parameter declarations, the names of
3577 its type parameters are types:
3580 fn to_vec<A: Clone>(xs: &[A]) -> Vec<A> {
3584 let first: A = xs[0].clone();
3585 let mut rest: Vec<A> = to_vec(&xs[1..]);
3586 rest.insert(0, first);
3591 Here, `first` has type `A`, referring to `to_vec`'s `A` type parameter; and `rest`
3592 has type `Vec<A>`, a vector with element type `A`.
3596 The special type `Self` has a meaning within traits and impls. In a trait definition, it refers
3597 to an implicit type parameter representing the "implementing" type. In an impl,
3598 it is an alias for the implementing type. For example, in:
3602 fn make_string(&self) -> String;
3605 impl Printable for String {
3606 fn make_string(&self) -> String {
3612 The notation `&self` is a shorthand for `self: &Self`. In this case,
3613 in the impl, `Self` refers to the value of type `String` that is the
3614 receiver for a call to the method `make_string`.
3618 Subtyping is implicit and can occur at any stage in type checking or
3619 inference. Subtyping in Rust is very restricted and occurs only due to
3620 variance with respect to lifetimes and between types with higher ranked
3621 lifetimes. If we were to erase lifetimes from types, then the only subtyping
3622 would be due to type equality.
3624 Consider the following example: string literals always have `'static`
3625 lifetime. Nevertheless, we can assign `s` to `t`:
3629 let s: &'static str = "hi";
3633 Since `'static` "lives longer" than `'a`, `&'static str` is a subtype of
3638 Coercions are defined in [RFC401]. A coercion is implicit and has no syntax.
3640 [RFC401]: https://github.com/rust-lang/rfcs/blob/master/text/0401-coercions.md
3644 A coercion can only occur at certain coercion sites in a program; these are
3645 typically places where the desired type is explicit or can be derived by
3646 propagation from explicit types (without type inference). Possible coercion
3649 * `let` statements where an explicit type is given.
3651 For example, `128` is coerced to have type `i8` in the following:
3657 * `static` and `const` statements (similar to `let` statements).
3659 * Arguments for function calls
3661 The value being coerced is the actual parameter, and it is coerced to
3662 the type of the formal parameter.
3664 For example, `128` is coerced to have type `i8` in the following:
3674 * Instantiations of struct or variant fields
3676 For example, `128` is coerced to have type `i8` in the following:
3679 struct Foo { x: i8 }
3686 * Function results, either the final line of a block if it is not
3687 semicolon-terminated or any expression in a `return` statement
3689 For example, `128` is coerced to have type `i8` in the following:
3697 If the expression in one of these coercion sites is a coercion-propagating
3698 expression, then the relevant sub-expressions in that expression are also
3699 coercion sites. Propagation recurses from these new coercion sites.
3700 Propagating expressions and their relevant sub-expressions are:
3702 * Array literals, where the array has type `[U; n]`. Each sub-expression in
3703 the array literal is a coercion site for coercion to type `U`.
3705 * Array literals with repeating syntax, where the array has type `[U; n]`. The
3706 repeated sub-expression is a coercion site for coercion to type `U`.
3708 * Tuples, where a tuple is a coercion site to type `(U_0, U_1, ..., U_n)`.
3709 Each sub-expression is a coercion site to the respective type, e.g. the
3710 zeroth sub-expression is a coercion site to type `U_0`.
3712 * Parenthesized sub-expressions (`(e)`): if the expression has type `U`, then
3713 the sub-expression is a coercion site to `U`.
3715 * Blocks: if a block has type `U`, then the last expression in the block (if
3716 it is not semicolon-terminated) is a coercion site to `U`. This includes
3717 blocks which are part of control flow statements, such as `if`/`else`, if
3718 the block has a known type.
3722 Coercion is allowed between the following types:
3724 * `T` to `U` if `T` is a subtype of `U` (*reflexive case*)
3726 * `T_1` to `T_3` where `T_1` coerces to `T_2` and `T_2` coerces to `T_3`
3729 Note that this is not fully supported yet
3733 * `*mut T` to `*const T`
3735 * `&T` to `*const T`
3737 * `&mut T` to `*mut T`
3739 * `&T` to `&U` if `T` implements `Deref<Target = U>`. For example:
3742 use std::ops::Deref;
3744 struct CharContainer {
3748 impl Deref for CharContainer {
3751 fn deref<'a>(&'a self) -> &'a char {
3756 fn foo(arg: &char) {}
3759 let x = &mut CharContainer { value: 'y' };
3760 foo(x); //&mut CharContainer is coerced to &char.
3764 * `&mut T` to `&mut U` if `T` implements `DerefMut<Target = U>`.
3766 * TyCtor(`T`) to TyCtor(coerce_inner(`T`)), where TyCtor(`T`) is one of
3774 - coerce_inner(`[T, ..n]`) = `[T]`
3775 - coerce_inner(`T`) = `U` where `T` is a concrete type which implements the
3778 In the future, coerce_inner will be recursively extended to tuples and
3779 structs. In addition, coercions from sub-traits to super-traits will be
3780 added. See [RFC401] for more details.
3784 Several traits define special evaluation behavior.
3788 The `Copy` trait changes the semantics of a type implementing it. Values whose
3789 type implements `Copy` are copied rather than moved upon assignment.
3791 ## The `Sized` trait
3793 The `Sized` trait indicates that the size of this type is known at compile-time.
3797 The `Drop` trait provides a destructor, to be run whenever a value of this type
3800 ## The `Deref` trait
3802 The `Deref<Target = U>` trait allows a type to implicitly implement all the methods
3803 of the type `U`. When attempting to resolve a method call, the compiler will search
3804 the top-level type for the implementation of the called method. If no such method is
3805 found, `.deref()` is called and the compiler continues to search for the method
3806 implementation in the returned type `U`.
3810 A Rust program's memory consists of a static set of *items* and a *heap*.
3811 Immutable portions of the heap may be safely shared between threads, mutable
3812 portions may not be safely shared, but several mechanisms for effectively-safe
3813 sharing of mutable values, built on unsafe code but enforcing a safe locking
3814 discipline, exist in the standard library.
3816 Allocations in the stack consist of *variables*, and allocations in the heap
3819 ### Memory allocation and lifetime
3821 The _items_ of a program are those functions, modules and types that have their
3822 value calculated at compile-time and stored uniquely in the memory image of the
3823 rust process. Items are neither dynamically allocated nor freed.
3825 The _heap_ is a general term that describes boxes. The lifetime of an
3826 allocation in the heap depends on the lifetime of the box values pointing to
3827 it. Since box values may themselves be passed in and out of frames, or stored
3828 in the heap, heap allocations may outlive the frame they are allocated within.
3830 ### Memory ownership
3832 When a stack frame is exited, its local allocations are all released, and its
3833 references to boxes are dropped.
3837 A _variable_ is a component of a stack frame, either a named function parameter,
3838 an anonymous [temporary](#lvalues-rvalues-and-temporaries), or a named local
3841 A _local variable_ (or *stack-local* allocation) holds a value directly,
3842 allocated within the stack's memory. The value is a part of the stack frame.
3844 Local variables are immutable unless declared otherwise like: `let mut x = ...`.
3846 Function parameters are immutable unless declared with `mut`. The `mut` keyword
3847 applies only to the following parameter (so `|mut x, y|` and `fn f(mut x:
3848 Box<i32>, y: Box<i32>)` declare one mutable variable `x` and one immutable
3851 Methods that take either `self` or `Box<Self>` can optionally place them in a
3852 mutable variable by prefixing them with `mut` (similar to regular arguments):
3856 fn change(mut self) -> Self;
3857 fn modify(mut self: Box<Self>) -> Box<Self>;
3861 Local variables are not initialized when allocated; the entire frame worth of
3862 local variables are allocated at once, on frame-entry, in an uninitialized
3863 state. Subsequent statements within a function may or may not initialize the
3864 local variables. Local variables can be used only after they have been
3865 initialized; this is enforced by the compiler.
3869 The Rust compiler supports various methods to link crates together both
3870 statically and dynamically. This section will explore the various methods to
3871 link Rust crates together, and more information about native libraries can be
3872 found in the [FFI section of the book][ffi].
3874 In one session of compilation, the compiler can generate multiple artifacts
3875 through the usage of either command line flags or the `crate_type` attribute.
3876 If one or more command line flags are specified, all `crate_type` attributes will
3877 be ignored in favor of only building the artifacts specified by command line.
3879 * `--crate-type=bin`, `#[crate_type = "bin"]` - A runnable executable will be
3880 produced. This requires that there is a `main` function in the crate which
3881 will be run when the program begins executing. This will link in all Rust and
3882 native dependencies, producing a distributable binary.
3884 * `--crate-type=lib`, `#[crate_type = "lib"]` - A Rust library will be produced.
3885 This is an ambiguous concept as to what exactly is produced because a library
3886 can manifest itself in several forms. The purpose of this generic `lib` option
3887 is to generate the "compiler recommended" style of library. The output library
3888 will always be usable by rustc, but the actual type of library may change from
3889 time-to-time. The remaining output types are all different flavors of
3890 libraries, and the `lib` type can be seen as an alias for one of them (but the
3891 actual one is compiler-defined).
3893 * `--crate-type=dylib`, `#[crate_type = "dylib"]` - A dynamic Rust library will
3894 be produced. This is different from the `lib` output type in that this forces
3895 dynamic library generation. The resulting dynamic library can be used as a
3896 dependency for other libraries and/or executables. This output type will
3897 create `*.so` files on linux, `*.dylib` files on osx, and `*.dll` files on
3900 * `--crate-type=staticlib`, `#[crate_type = "staticlib"]` - A static system
3901 library will be produced. This is different from other library outputs in that
3902 the Rust compiler will never attempt to link to `staticlib` outputs. The
3903 purpose of this output type is to create a static library containing all of
3904 the local crate's code along with all upstream dependencies. The static
3905 library is actually a `*.a` archive on linux and osx and a `*.lib` file on
3906 windows. This format is recommended for use in situations such as linking
3907 Rust code into an existing non-Rust application because it will not have
3908 dynamic dependencies on other Rust code.
3910 * `--crate-type=rlib`, `#[crate_type = "rlib"]` - A "Rust library" file will be
3911 produced. This is used as an intermediate artifact and can be thought of as a
3912 "static Rust library". These `rlib` files, unlike `staticlib` files, are
3913 interpreted by the Rust compiler in future linkage. This essentially means
3914 that `rustc` will look for metadata in `rlib` files like it looks for metadata
3915 in dynamic libraries. This form of output is used to produce statically linked
3916 executables as well as `staticlib` outputs.
3918 Note that these outputs are stackable in the sense that if multiple are
3919 specified, then the compiler will produce each form of output at once without
3920 having to recompile. However, this only applies for outputs specified by the
3921 same method. If only `crate_type` attributes are specified, then they will all
3922 be built, but if one or more `--crate-type` command line flags are specified,
3923 then only those outputs will be built.
3925 With all these different kinds of outputs, if crate A depends on crate B, then
3926 the compiler could find B in various different forms throughout the system. The
3927 only forms looked for by the compiler, however, are the `rlib` format and the
3928 dynamic library format. With these two options for a dependent library, the
3929 compiler must at some point make a choice between these two formats. With this
3930 in mind, the compiler follows these rules when determining what format of
3931 dependencies will be used:
3933 1. If a static library is being produced, all upstream dependencies are
3934 required to be available in `rlib` formats. This requirement stems from the
3935 reason that a dynamic library cannot be converted into a static format.
3937 Note that it is impossible to link in native dynamic dependencies to a static
3938 library, and in this case warnings will be printed about all unlinked native
3939 dynamic dependencies.
3941 2. If an `rlib` file is being produced, then there are no restrictions on what
3942 format the upstream dependencies are available in. It is simply required that
3943 all upstream dependencies be available for reading metadata from.
3945 The reason for this is that `rlib` files do not contain any of their upstream
3946 dependencies. It wouldn't be very efficient for all `rlib` files to contain a
3947 copy of `libstd.rlib`!
3949 3. If an executable is being produced and the `-C prefer-dynamic` flag is not
3950 specified, then dependencies are first attempted to be found in the `rlib`
3951 format. If some dependencies are not available in an rlib format, then
3952 dynamic linking is attempted (see below).
3954 4. If a dynamic library or an executable that is being dynamically linked is
3955 being produced, then the compiler will attempt to reconcile the available
3956 dependencies in either the rlib or dylib format to create a final product.
3958 A major goal of the compiler is to ensure that a library never appears more
3959 than once in any artifact. For example, if dynamic libraries B and C were
3960 each statically linked to library A, then a crate could not link to B and C
3961 together because there would be two copies of A. The compiler allows mixing
3962 the rlib and dylib formats, but this restriction must be satisfied.
3964 The compiler currently implements no method of hinting what format a library
3965 should be linked with. When dynamically linking, the compiler will attempt to
3966 maximize dynamic dependencies while still allowing some dependencies to be
3967 linked in via an rlib.
3969 For most situations, having all libraries available as a dylib is recommended
3970 if dynamically linking. For other situations, the compiler will emit a
3971 warning if it is unable to determine which formats to link each library with.
3973 In general, `--crate-type=bin` or `--crate-type=lib` should be sufficient for
3974 all compilation needs, and the other options are just available if more
3975 fine-grained control is desired over the output format of a Rust crate.
3979 Unsafe operations are those that potentially violate the memory-safety
3980 guarantees of Rust's static semantics.
3982 The following language level features cannot be used in the safe subset of
3985 - Dereferencing a [raw pointer](#pointer-types).
3986 - Reading or writing a [mutable static variable](#mutable-statics).
3987 - Calling an unsafe function (including an intrinsic or foreign function).
3991 Unsafe functions are functions that are not safe in all contexts and/or for all
3992 possible inputs. Such a function must be prefixed with the keyword `unsafe` and
3993 can only be called from an `unsafe` block or another `unsafe` function.
3997 A block of code can be prefixed with the `unsafe` keyword, to permit calling
3998 `unsafe` functions or dereferencing raw pointers within a safe function.
4000 When a programmer has sufficient conviction that a sequence of potentially
4001 unsafe operations is actually safe, they can encapsulate that sequence (taken
4002 as a whole) within an `unsafe` block. The compiler will consider uses of such
4003 code safe, in the surrounding context.
4005 Unsafe blocks are used to wrap foreign libraries, make direct use of hardware
4006 or implement features not directly present in the language. For example, Rust
4007 provides the language features necessary to implement memory-safe concurrency
4008 in the language but the implementation of threads and message passing is in the
4011 Rust's type system is a conservative approximation of the dynamic safety
4012 requirements, so in some cases there is a performance cost to using safe code.
4013 For example, a doubly-linked list is not a tree structure and can only be
4014 represented with reference-counted pointers in safe code. By using `unsafe`
4015 blocks to represent the reverse links as raw pointers, it can be implemented
4018 ## Behavior considered undefined
4020 The following is a list of behavior which is forbidden in all Rust code,
4021 including within `unsafe` blocks and `unsafe` functions. Type checking provides
4022 the guarantee that these issues are never caused by safe code.
4025 * Dereferencing a null/dangling raw pointer
4026 * Reads of [undef](http://llvm.org/docs/LangRef.html#undefined-values)
4027 (uninitialized) memory
4028 * Breaking the [pointer aliasing
4029 rules](http://llvm.org/docs/LangRef.html#pointer-aliasing-rules)
4030 with raw pointers (a subset of the rules used by C)
4031 * `&mut` and `&` follow LLVM’s scoped [noalias] model, except if the `&T`
4032 contains an `UnsafeCell<U>`. Unsafe code must not violate these aliasing
4034 * Mutating non-mutable data (that is, data reached through a shared reference or
4035 data owned by a `let` binding), unless that data is contained within an `UnsafeCell<U>`.
4036 * Invoking undefined behavior via compiler intrinsics:
4037 * Indexing outside of the bounds of an object with `std::ptr::offset`
4038 (`offset` intrinsic), with
4039 the exception of one byte past the end which is permitted.
4040 * Using `std::ptr::copy_nonoverlapping_memory` (`memcpy32`/`memcpy64`
4041 intrinsics) on overlapping buffers
4042 * Invalid values in primitive types, even in private fields/locals:
4043 * Dangling/null references or boxes
4044 * A value other than `false` (0) or `true` (1) in a `bool`
4045 * A discriminant in an `enum` not included in the type definition
4046 * A value in a `char` which is a surrogate or above `char::MAX`
4047 * Non-UTF-8 byte sequences in a `str`
4048 * Unwinding into Rust from foreign code or unwinding from Rust into foreign
4049 code. Rust's failure system is not compatible with exception handling in
4050 other languages. Unwinding must be caught and handled at FFI boundaries.
4052 [noalias]: http://llvm.org/docs/LangRef.html#noalias
4054 ## Behavior not considered unsafe
4056 This is a list of behavior not considered *unsafe* in Rust terms, but that may
4060 * Leaks of memory and other resources
4061 * Exiting without calling destructors
4063 - Overflow is considered "unexpected" behavior and is always user-error,
4064 unless the `wrapping` primitives are used. In non-optimized builds, the compiler
4065 will insert debug checks that panic on overflow, but in optimized builds overflow
4066 instead results in wrapped values. See [RFC 560] for the rationale and more details.
4068 [RFC 560]: https://github.com/rust-lang/rfcs/blob/master/text/0560-integer-overflow.md
4070 # Appendix: Influences
4072 Rust is not a particularly original language, with design elements coming from
4073 a wide range of sources. Some of these are listed below (including elements
4074 that have since been removed):
4076 * SML, OCaml: algebraic data types, pattern matching, type inference,
4077 semicolon statement separation
4078 * C++: references, RAII, smart pointers, move semantics, monomorphization,
4080 * ML Kit, Cyclone: region based memory management
4081 * Haskell (GHC): typeclasses, type families
4082 * Newsqueak, Alef, Limbo: channels, concurrency
4083 * Erlang: message passing, thread failure, ~~linked thread failure~~,
4084 ~~lightweight concurrency~~
4085 * Swift: optional bindings
4086 * Scheme: hygienic macros
4088 * Ruby: ~~block syntax~~
4089 * NIL, Hermes: ~~typestate~~
4090 * [Unicode Annex #31](http://www.unicode.org/reports/tr31/): identifier and
4093 [ffi]: book/ffi.html
4094 [plugin]: book/compiler-plugins.html