5 This document is the primary reference for the Rust programming language. It
6 provides three kinds of material:
8 - Chapters that formally define the language grammar and, for each
9 construct, informally describe its semantics and give examples of its
11 - Chapters that informally describe the memory model, concurrency model,
12 runtime services, linkage model and debugging facilities.
13 - Appendix chapters providing rationale and references to languages that
14 influenced the design.
16 This document does not serve as an introduction to the language. Background
17 familiarity with the language is assumed. A separate [guide] is available to
18 help acquire such background familiarity.
20 This document also does not serve as a reference to the [standard] library
21 included in the language distribution. Those libraries are documented
22 separately by extracting documentation attributes from their source code. Many
23 of the features that one might expect to be language features are library
24 features in Rust, so what you're looking for may be there, not here.
27 [standard]: std/index.html
31 Rust's grammar is defined over Unicode codepoints, each conventionally denoted
32 `U+XXXX`, for 4 or more hexadecimal digits `X`. _Most_ of Rust's grammar is
33 confined to the ASCII range of Unicode, and is described in this document by a
34 dialect of Extended Backus-Naur Form (EBNF), specifically a dialect of EBNF
35 supported by common automated LL(k) parsing tools such as `llgen`, rather than
36 the dialect given in ISO 14977. The dialect can be defined self-referentially
41 rule : nonterminal ':' productionrule ';' ;
42 productionrule : production [ '|' production ] * ;
44 term : element repeats ;
45 element : LITERAL | IDENTIFIER | '[' productionrule ']' ;
46 repeats : [ '*' | '+' ] NUMBER ? | NUMBER ? | '?' ;
51 - Whitespace in the grammar is ignored.
52 - Square brackets are used to group rules.
53 - `LITERAL` is a single printable ASCII character, or an escaped hexadecimal
54 ASCII code of the form `\xQQ`, in single quotes, denoting the corresponding
55 Unicode codepoint `U+00QQ`.
56 - `IDENTIFIER` is a nonempty string of ASCII letters and underscores.
57 - The `repeat` forms apply to the adjacent `element`, and are as follows:
58 - `?` means zero or one repetition
59 - `*` means zero or more repetitions
60 - `+` means one or more repetitions
61 - NUMBER trailing a repeat symbol gives a maximum repetition count
62 - NUMBER on its own gives an exact repetition count
64 This EBNF dialect should hopefully be familiar to many readers.
66 ## Unicode productions
68 A few productions in Rust's grammar permit Unicode codepoints outside the ASCII
69 range. We define these productions in terms of character properties specified
70 in the Unicode standard, rather than in terms of ASCII-range codepoints. The
71 section [Special Unicode Productions](#special-unicode-productions) lists these
74 ## String table productions
76 Some rules in the grammar — notably [unary
77 operators](#unary-operator-expressions), [binary
78 operators](#binary-operator-expressions), and [keywords](#keywords) — are
79 given in a simplified form: as a listing of a table of unquoted, printable
80 whitespace-separated strings. These cases form a subset of the rules regarding
81 the [token](#tokens) rule, and are assumed to be the result of a
82 lexical-analysis phase feeding the parser, driven by a DFA, operating over the
83 disjunction of all such string table entries.
85 When such a string enclosed in double-quotes (`"`) occurs inside the grammar,
86 it is an implicit reference to a single member of such a string table
87 production. See [tokens](#tokens) for more information.
93 Rust input is interpreted as a sequence of Unicode codepoints encoded in UTF-8.
94 Most Rust grammar rules are defined in terms of printable ASCII-range
95 codepoints, but a small number are defined in terms of Unicode properties or
96 explicit codepoint lists. [^inputformat]
98 [^inputformat]: Substitute definitions for the special Unicode productions are
99 provided to the grammar verifier, restricted to ASCII range, when verifying the
100 grammar in this document.
102 ## Special Unicode Productions
104 The following productions in the Rust grammar are defined in terms of Unicode
105 properties: `ident`, `non_null`, `non_star`, `non_eol`, `non_slash_or_star`,
106 `non_single_quote` and `non_double_quote`.
110 The `ident` production is any nonempty Unicode string of the following form:
112 - The first character has property `XID_start`
113 - The remaining characters have property `XID_continue`
115 that does _not_ occur in the set of [keywords](#keywords).
117 > **Note**: `XID_start` and `XID_continue` as character properties cover the
118 > character ranges used to form the more familiar C and Java language-family
121 ### Delimiter-restricted productions
123 Some productions are defined by exclusion of particular Unicode characters:
125 - `non_null` is any single Unicode character aside from `U+0000` (null)
126 - `non_eol` is `non_null` restricted to exclude `U+000A` (`'\n'`)
127 - `non_star` is `non_null` restricted to exclude `U+002A` (`*`)
128 - `non_slash_or_star` is `non_null` restricted to exclude `U+002F` (`/`) and `U+002A` (`*`)
129 - `non_single_quote` is `non_null` restricted to exclude `U+0027` (`'`)
130 - `non_double_quote` is `non_null` restricted to exclude `U+0022` (`"`)
135 comment : block_comment | line_comment ;
136 block_comment : "/*" block_comment_body * "*/" ;
137 block_comment_body : [block_comment | character] * ;
138 line_comment : "//" non_eol * ;
141 Comments in Rust code follow the general C++ style of line and block-comment
142 forms. Nested block comments are supported.
144 Line comments beginning with exactly _three_ slashes (`///`), and block
145 comments beginning with exactly one repeated asterisk in the block-open
146 sequence (`/**`), are interpreted as a special syntax for `doc`
147 [attributes](#attributes). That is, they are equivalent to writing
148 `#[doc="..."]` around the body of the comment (this includes the comment
149 characters themselves, ie `/// Foo` turns into `#[doc="/// Foo"]`).
151 `//!` comments apply to the parent of the comment, rather than the item that
152 follows. `//!` comments are usually used to display information on the crate
155 Non-doc comments are interpreted as a form of whitespace.
160 whitespace_char : '\x20' | '\x09' | '\x0a' | '\x0d' ;
161 whitespace : [ whitespace_char | comment ] + ;
164 The `whitespace_char` production is any nonempty Unicode string consisting of
165 any of the following Unicode characters: `U+0020` (space, `' '`), `U+0009`
166 (tab, `'\t'`), `U+000A` (LF, `'\n'`), `U+000D` (CR, `'\r'`).
168 Rust is a "free-form" language, meaning that all forms of whitespace serve only
169 to separate _tokens_ in the grammar, and have no semantic significance.
171 A Rust program has identical meaning if each whitespace element is replaced
172 with any other legal whitespace element, such as a single space character.
177 simple_token : keyword | unop | binop ;
178 token : simple_token | ident | literal | symbol | whitespace token ;
181 Tokens are primitive productions in the grammar defined by regular
182 (non-recursive) languages. "Simple" tokens are given in [string table
183 production](#string-table-productions) form, and occur in the rest of the
184 grammar as double-quoted strings. Other tokens have exact rules given.
188 <p id="keyword-table-marker"></p>
191 |----------|----------|----------|----------|---------|
192 | abstract | alignof | as | be | box |
193 | break | const | continue | crate | do |
194 | else | enum | extern | false | final |
195 | fn | for | if | impl | in |
196 | let | loop | match | mod | move |
197 | mut | offsetof | override | priv | pub |
198 | pure | ref | return | sizeof | static |
199 | self | struct | super | true | trait |
200 | type | typeof | unsafe | unsized | use |
201 | virtual | where | while | yield |
204 Each of these keywords has special meaning in its grammar, and all of them are
205 excluded from the `ident` rule.
207 Note that some of these keywords are reserved, and do not currently do
212 A literal is an expression consisting of a single token, rather than a sequence
213 of tokens, that immediately and directly denotes the value it evaluates to,
214 rather than referring to it by name or some other evaluation rule. A literal is
215 a form of constant expression, so is evaluated (primarily) at compile time.
219 literal : [ string_lit | char_lit | byte_string_lit | byte_lit | num_lit ] lit_suffix ?;
222 The optional suffix is only used for certain numeric literals, but is
223 reserved for future extension, that is, the above gives the lexical
224 grammar, but a Rust parser will reject everything but the 12 special
225 cases mentioned in [Number literals](#number-literals) below.
229 ##### Characters and strings
231 | | Example | Number of `#` pairs allowed | Available characters | Escapes | Equivalent to |
232 |---|---------|-----------------------------|----------------------|---------|---------------|
233 | [Character](#character-literals) | `'H'` | `N/A` | All unicode | `\'` & [Byte escapes](#byte-escapes) & [Unicode escapes](#unicode-escapes) | `N/A` |
234 | [String](#string-literals) | `"hello"` | `N/A` | All unicode | `\"` & [Byte escapes](#byte-escapes) & [Unicode escapes](#unicode-escapes) | `N/A` |
235 | [Raw](#raw-string-literals) | `r##"hello"##` | `0...` | All unicode | `N/A` | `N/A` |
236 | [Byte](#byte-literals) | `b'H'` | `N/A` | All ASCII | `\'` & [Byte escapes](#byte-escapes) | `u8` |
237 | [Byte string](#byte-string-literals) | `b"hello"` | `N/A` | All ASCII | `\"` & [Byte escapes](#byte-escapes) | `&'static [u8]` |
238 | [Raw byte string](#raw-byte-string-literals) | `br##"hello"##` | `0...` | All ASCII | `N/A` | `&'static [u8]` (unsure...not stated) |
244 | `\x7F` | 8-bit character code (exactly 2 digits) |
246 | `\r` | Carriage return |
250 ##### Unicode escapes
253 | `\u7FFF` | 16-bit character code (exactly 4 digits) |
254 | `\U7EEEFFFF` | 32-bit character code (exactly 8 digits) |
258 | [Number literals](#number-literals)`*` | Example | Exponentiation | Suffixes |
259 |----------------------------------------|---------|----------------|----------|
260 | Decimal integer | `98_222i` | `N/A` | Integer suffixes |
261 | Hex integer | `0xffi` | `N/A` | Integer suffixes |
262 | Octal integer | `0o77i` | `N/A` | Integer suffixes |
263 | Binary integer | `0b1111_0000i` | `N/A` | Integer suffixes |
264 | Floating-point | `123.0E+77f64` | `Optional` | Floating-point suffixes |
266 `*` All number literals allow `_` as a visual separator: `1_234.0E+18f64`
269 | Integer | Floating-point |
270 |---------|----------------|
271 | `i` (`int`), `u` (`uint`), `u8`, `i8`, `u16`, `i16`, `u32`, `i32`, `u64`, `i64` | `f32`, `f64` |
273 #### Character and string literals
276 char_lit : '\x27' char_body '\x27' ;
277 string_lit : '"' string_body * '"' | 'r' raw_string ;
279 char_body : non_single_quote
280 | '\x5c' [ '\x27' | common_escape | unicode_escape ] ;
282 string_body : non_double_quote
283 | '\x5c' [ '\x22' | common_escape | unicode_escape ] ;
284 raw_string : '"' raw_string_body '"' | '#' raw_string '#' ;
286 common_escape : '\x5c'
287 | 'n' | 'r' | 't' | '0'
289 unicode_escape : 'u' hex_digit 4
292 hex_digit : 'a' | 'b' | 'c' | 'd' | 'e' | 'f'
293 | 'A' | 'B' | 'C' | 'D' | 'E' | 'F'
295 oct_digit : '0' | '1' | '2' | '3' | '4' | '5' | '6' | '7' ;
296 dec_digit : '0' | nonzero_dec ;
297 nonzero_dec: '1' | '2' | '3' | '4'
298 | '5' | '6' | '7' | '8' | '9' ;
301 ##### Character literals
303 A _character literal_ is a single Unicode character enclosed within two
304 `U+0027` (single-quote) characters, with the exception of `U+0027` itself,
305 which must be _escaped_ by a preceding U+005C character (`\`).
307 ##### String literals
309 A _string literal_ is a sequence of any Unicode characters enclosed within two
310 `U+0022` (double-quote) characters, with the exception of `U+0022` itself,
311 which must be _escaped_ by a preceding `U+005C` character (`\`), or a _raw
314 ##### Character escapes
316 Some additional _escapes_ are available in either character or non-raw string
317 literals. An escape starts with a `U+005C` (`\`) and continues with one of the
320 * An _8-bit codepoint escape_ escape starts with `U+0078` (`x`) and is
321 followed by exactly two _hex digits_. It denotes the Unicode codepoint
322 equal to the provided hex value.
323 * A _16-bit codepoint escape_ starts with `U+0075` (`u`) and is followed
324 by exactly four _hex digits_. It denotes the Unicode codepoint equal to
325 the provided hex value.
326 * A _32-bit codepoint escape_ starts with `U+0055` (`U`) and is followed
327 by exactly eight _hex digits_. It denotes the Unicode codepoint equal to
328 the provided hex value.
329 * A _whitespace escape_ is one of the characters `U+006E` (`n`), `U+0072`
330 (`r`), or `U+0074` (`t`), denoting the unicode values `U+000A` (LF),
331 `U+000D` (CR) or `U+0009` (HT) respectively.
332 * The _backslash escape_ is the character `U+005C` (`\`) which must be
333 escaped in order to denote *itself*.
335 ##### Raw string literals
337 Raw string literals do not process any escapes. They start with the character
338 `U+0072` (`r`), followed by zero or more of the character `U+0023` (`#`) and a
339 `U+0022` (double-quote) character. The _raw string body_ is not defined in the
340 EBNF grammar above: it can contain any sequence of Unicode characters and is
341 terminated only by another `U+0022` (double-quote) character, followed by the
342 same number of `U+0023` (`#`) characters that preceded the opening `U+0022`
343 (double-quote) character.
345 All Unicode characters contained in the raw string body represent themselves,
346 the characters `U+0022` (double-quote) (except when followed by at least as
347 many `U+0023` (`#`) characters as were used to start the raw string literal) or
348 `U+005C` (`\`) do not have any special meaning.
350 Examples for string literals:
353 "foo"; r"foo"; // foo
354 "\"foo\""; r#""foo""#; // "foo"
357 r##"foo #"# bar"##; // foo #"# bar
359 "\x52"; "R"; r"R"; // R
360 "\\x52"; r"\x52"; // \x52
363 #### Byte and byte string literals
366 byte_lit : "b\x27" byte_body '\x27' ;
367 byte_string_lit : "b\x22" string_body * '\x22' | "br" raw_byte_string ;
369 byte_body : ascii_non_single_quote
370 | '\x5c' [ '\x27' | common_escape ] ;
372 byte_string_body : ascii_non_double_quote
373 | '\x5c' [ '\x22' | common_escape ] ;
374 raw_byte_string : '"' raw_byte_string_body '"' | '#' raw_byte_string '#' ;
380 A _byte literal_ is a single ASCII character (in the `U+0000` to `U+007F`
381 range) enclosed within two `U+0027` (single-quote) characters, with the
382 exception of `U+0027` itself, which must be _escaped_ by a preceding U+005C
383 character (`\`), or a single _escape_. It is equivalent to a `u8` unsigned
384 8-bit integer _number literal_.
386 ##### Byte string literals
388 A _byte string literal_ is a sequence of ASCII characters and _escapes_
389 enclosed within two `U+0022` (double-quote) characters, with the exception of
390 `U+0022` itself, which must be _escaped_ by a preceding `U+005C` character
391 (`\`), or a _raw byte string literal_. It is equivalent to a `&'static [u8]`
392 borrowed array of unsigned 8-bit integers.
394 Some additional _escapes_ are available in either byte or non-raw byte string
395 literals. An escape starts with a `U+005C` (`\`) and continues with one of the
398 * An _byte escape_ escape starts with `U+0078` (`x`) and is
399 followed by exactly two _hex digits_. It denotes the byte
400 equal to the provided hex value.
401 * A _whitespace escape_ is one of the characters `U+006E` (`n`), `U+0072`
402 (`r`), or `U+0074` (`t`), denoting the bytes values `0x0A` (ASCII LF),
403 `0x0D` (ASCII CR) or `0x09` (ASCII HT) respectively.
404 * The _backslash escape_ is the character `U+005C` (`\`) which must be
405 escaped in order to denote its ASCII encoding `0x5C`.
407 ##### Raw byte string literals
409 Raw byte string literals do not process any escapes. They start with the
410 character `U+0062` (`b`), followed by `U+0072` (`r`), followed by zero or more
411 of the character `U+0023` (`#`), and a `U+0022` (double-quote) character. The
412 _raw string body_ is not defined in the EBNF grammar above: it can contain any
413 sequence of ASCII characters and is terminated only by another `U+0022`
414 (double-quote) character, followed by the same number of `U+0023` (`#`)
415 characters that preceded the opening `U+0022` (double-quote) character. A raw
416 byte string literal can not contain any non-ASCII byte.
418 All characters contained in the raw string body represent their ASCII encoding,
419 the characters `U+0022` (double-quote) (except when followed by at least as
420 many `U+0023` (`#`) characters as were used to start the raw string literal) or
421 `U+005C` (`\`) do not have any special meaning.
423 Examples for byte string literals:
426 b"foo"; br"foo"; // foo
427 b"\"foo\""; br#""foo""#; // "foo"
430 br##"foo #"# bar"##; // foo #"# bar
432 b"\x52"; b"R"; br"R"; // R
433 b"\\x52"; br"\x52"; // \x52
439 num_lit : nonzero_dec [ dec_digit | '_' ] * float_suffix ?
440 | '0' [ [ dec_digit | '_' ] * float_suffix ?
441 | 'b' [ '1' | '0' | '_' ] +
442 | 'o' [ oct_digit | '_' ] +
443 | 'x' [ hex_digit | '_' ] + ] ;
445 float_suffix : [ exponent | '.' dec_lit exponent ? ] ? ;
447 exponent : ['E' | 'e'] ['-' | '+' ] ? dec_lit ;
448 dec_lit : [ dec_digit | '_' ] + ;
451 A _number literal_ is either an _integer literal_ or a _floating-point
452 literal_. The grammar for recognizing the two kinds of literals is mixed.
454 ##### Integer literals
456 An _integer literal_ has one of four forms:
458 * A _decimal literal_ starts with a *decimal digit* and continues with any
459 mixture of *decimal digits* and _underscores_.
460 * A _hex literal_ starts with the character sequence `U+0030` `U+0078`
461 (`0x`) and continues as any mixture of hex digits and underscores.
462 * An _octal literal_ starts with the character sequence `U+0030` `U+006F`
463 (`0o`) and continues as any mixture of octal digits and underscores.
464 * A _binary literal_ starts with the character sequence `U+0030` `U+0062`
465 (`0b`) and continues as any mixture of binary digits and underscores.
467 Like any literal, an integer literal may be followed (immediately,
468 without any spaces) by an _integer suffix_, which forcibly sets the
469 type of the literal. There are 10 valid values for an integer suffix:
471 * The `i` and `u` suffixes give the literal type `int` or `uint`,
473 * Each of the signed and unsigned machine types `u8`, `i8`,
474 `u16`, `i16`, `u32`, `i32`, `u64` and `i64`
475 give the literal the corresponding machine type.
477 The type of an _unsuffixed_ integer literal is determined by type inference.
478 If an integer type can be _uniquely_ determined from the surrounding program
479 context, the unsuffixed integer literal has that type. If the program context
480 underconstrains the type, it is considered a static type error; if the program
481 context overconstrains the type, it is also considered a static type error.
483 Examples of integer literals of various forms:
490 0o70_i16; // type i16
491 0b1111_1111_1001_0000_i32; // type i32
494 ##### Floating-point literals
496 A _floating-point literal_ has one of two forms:
498 * A _decimal literal_ followed by a period character `U+002E` (`.`). This is
499 optionally followed by another decimal literal, with an optional _exponent_.
500 * A single _decimal literal_ followed by an _exponent_.
502 By default, a floating-point literal has a generic type, and, like integer
503 literals, the type must be uniquely determined from the context. There are two valid
504 _floating-point suffixes_, `f32` and `f64` (the 32-bit and 64-bit floating point
505 types), which explicitly determine the type of the literal.
507 Examples of floating-point literals of various forms:
510 123.0f64; // type f64
513 12E+99_f64; // type f64
514 let x: f64 = 2.; // type f64
517 This last example is different because it is not possible to use the suffix
518 syntax with a floating point literal ending in a period. `2.f64` would attempt
519 to call a method named `f64` on `2`.
521 #### Boolean literals
523 The two values of the boolean type are written `true` and `false`.
529 | '#' | '[' | ']' | '(' | ')' | '{' | '}'
533 Symbols are a general class of printable [token](#tokens) that play structural
534 roles in a variety of grammar productions. They are catalogued here for
535 completeness as the set of remaining miscellaneous printable tokens that do not
536 otherwise appear as [unary operators](#unary-operator-expressions), [binary
537 operators](#binary-operator-expressions), or [keywords](#keywords).
543 expr_path : [ "::" ] ident [ "::" expr_path_tail ] + ;
544 expr_path_tail : '<' type_expr [ ',' type_expr ] + '>'
547 type_path : ident [ type_path_tail ] + ;
548 type_path_tail : '<' type_expr [ ',' type_expr ] + '>'
552 A _path_ is a sequence of one or more path components _logically_ separated by
553 a namespace qualifier (`::`). If a path consists of only one component, it may
554 refer to either an [item](#items) or a [slot](#memory-slots) in a local control
555 scope. If a path has multiple components, it refers to an item.
557 Every item has a _canonical path_ within its crate, but the path naming an item
558 is only meaningful within a given crate. There is no global namespace across
559 crates; an item's canonical path merely identifies it within the crate.
561 Two examples of simple paths consisting of only identifier components:
568 Path components are usually [identifiers](#identifiers), but the trailing
569 component of a path may be an angle-bracket-enclosed list of type arguments. In
570 [expression](#expressions) context, the type argument list is given after a
571 final (`::`) namespace qualifier in order to disambiguate it from a relational
572 expression involving the less-than symbol (`<`). In type expression context,
573 the final namespace qualifier is omitted.
575 Two examples of paths with type arguments:
578 # struct HashMap<K, V>;
580 # fn id<T>(t: T) -> T { t }
581 type T = HashMap<int,String>; // Type arguments used in a type expression
582 let x = id::<int>(10); // Type arguments used in a call expression
586 Paths can be denoted with various leading qualifiers to change the meaning of
589 * Paths starting with `::` are considered to be global paths where the
590 components of the path start being resolved from the crate root. Each
591 identifier in the path must resolve to an item.
599 ::a::foo(); // call a's foo function
605 * Paths starting with the keyword `super` begin resolution relative to the
606 parent module. Each further identifier must resolve to an item
614 super::a::foo(); // call a's foo function
620 * Paths starting with the keyword `self` begin resolution relative to the
621 current module. Each further identifier must resolve to an item.
633 A number of minor features of Rust are not central enough to have their own
634 syntax, and yet are not implementable as functions. Instead, they are given
635 names, and invoked through a consistent syntax: `name!(...)`. Examples include:
637 * `format!` : format data into a string
638 * `env!` : look up an environment variable's value at compile time
639 * `file!`: return the path to the file being compiled
640 * `stringify!` : pretty-print the Rust expression given as an argument
641 * `include!` : include the Rust expression in the given file
642 * `include_str!` : include the contents of the given file as a string
643 * `include_bytes!` : include the contents of the given file as a binary blob
644 * `error!`, `warn!`, `info!`, `debug!` : provide diagnostic information.
646 All of the above extensions are expressions with values.
648 Users of `rustc` can define new syntax extensions in two ways:
650 * [Compiler plugins](guide-plugin.html#syntax-extensions) can include arbitrary
651 Rust code that manipulates syntax trees at compile time.
653 * [Macros](guide-macros.html) define new syntax in a higher-level,
659 expr_macro_rules : "macro_rules" '!' ident '(' macro_rule * ')' ;
660 macro_rule : '(' matcher * ')' "=>" '(' transcriber * ')' ';' ;
661 matcher : '(' matcher * ')' | '[' matcher * ']'
662 | '{' matcher * '}' | '$' ident ':' ident
663 | '$' '(' matcher * ')' sep_token? [ '*' | '+' ]
664 | non_special_token ;
665 transcriber : '(' transcriber * ')' | '[' transcriber * ']'
666 | '{' transcriber * '}' | '$' ident
667 | '$' '(' transcriber * ')' sep_token? [ '*' | '+' ]
668 | non_special_token ;
671 User-defined syntax extensions are called "macros", and the `macro_rules`
672 syntax extension defines them. Currently, user-defined macros can expand to
673 expressions, statements, items, or patterns.
675 (A `sep_token` is any token other than `*` and `+`. A `non_special_token` is
676 any token other than a delimiter or `$`.)
678 The macro expander looks up macro invocations by name, and tries each macro
679 rule in turn. It transcribes the first successful match. Matching and
680 transcription are closely related to each other, and we will describe them
685 The macro expander matches and transcribes every token that does not begin with
686 a `$` literally, including delimiters. For parsing reasons, delimiters must be
687 balanced, but they are otherwise not special.
689 In the matcher, `$` _name_ `:` _designator_ matches the nonterminal in the Rust
690 syntax named by _designator_. Valid designators are `item`, `block`, `stmt`,
691 `pat`, `expr`, `ty` (type), `ident`, `path`, `matchers` (lhs of the `=>` in
692 macro rules), `tt` (rhs of the `=>` in macro rules). In the transcriber, the
693 designator is already known, and so only the name of a matched nonterminal
694 comes after the dollar sign.
696 In both the matcher and transcriber, the Kleene star-like operator indicates
697 repetition. The Kleene star operator consists of `$` and parens, optionally
698 followed by a separator token, followed by `*` or `+`. `*` means zero or more
699 repetitions, `+` means at least one repetition. The parens are not matched or
700 transcribed. On the matcher side, a name is bound to _all_ of the names it
701 matches, in a structure that mimics the structure of the repetition encountered
702 on a successful match. The job of the transcriber is to sort that structure
705 The rules for transcription of these repetitions are called "Macro By Example".
706 Essentially, one "layer" of repetition is discharged at a time, and all of them
707 must be discharged by the time a name is transcribed. Therefore, `( $( $i:ident
708 ),* ) => ( $i )` is an invalid macro, but `( $( $i:ident ),* ) => ( $( $i:ident
709 ),* )` is acceptable (if trivial).
711 When Macro By Example encounters a repetition, it examines all of the `$`
712 _name_ s that occur in its body. At the "current layer", they all must repeat
713 the same number of times, so ` ( $( $i:ident ),* ; $( $j:ident ),* ) => ( $(
714 ($i,$j) ),* )` is valid if given the argument `(a,b,c ; d,e,f)`, but not
715 `(a,b,c ; d,e)`. The repetition walks through the choices at that layer in
716 lockstep, so the former input transcribes to `( (a,d), (b,e), (c,f) )`.
718 Nested repetitions are allowed.
720 ### Parsing limitations
722 The parser used by the macro system is reasonably powerful, but the parsing of
723 Rust syntax is restricted in two ways:
725 1. The parser will always parse as much as possible. If it attempts to match
726 `$i:expr [ , ]` against `8 [ , ]`, it will attempt to parse `i` as an array
727 index operation and fail. Adding a separator can solve this problem.
728 2. The parser must have eliminated all ambiguity by the time it reaches a `$`
729 _name_ `:` _designator_. This requirement most often affects name-designator
730 pairs when they occur at the beginning of, or immediately after, a `$(...)*`;
731 requiring a distinctive token in front can solve the problem.
733 ## Syntax extensions useful for the macro author
735 * `log_syntax!` : print out the arguments at compile time
736 * `trace_macros!` : supply `true` or `false` to enable or disable macro expansion logging
737 * `stringify!` : turn the identifier argument into a string literal
738 * `concat!` : concatenates a comma-separated list of literals
739 * `concat_idents!` : create a new identifier by concatenating the arguments
741 # Crates and source files
743 Rust is a *compiled* language. Its semantics obey a *phase distinction*
744 between compile-time and run-time. Those semantic rules that have a *static
745 interpretation* govern the success or failure of compilation. We refer to
746 these rules as "static semantics". Semantic rules called "dynamic semantics"
747 govern the behavior of programs at run-time. A program that fails to compile
748 due to violation of a compile-time rule has no defined dynamic semantics; the
749 compiler should halt with an error report, and produce no executable artifact.
751 The compilation model centers on artifacts called _crates_. Each compilation
752 processes a single crate in source form, and if successful, produces a single
753 crate in binary form: either an executable or a library.[^cratesourcefile]
755 [^cratesourcefile]: A crate is somewhat analogous to an *assembly* in the
756 ECMA-335 CLI model, a *library* in the SML/NJ Compilation Manager, a *unit*
757 in the Owens and Flatt module system, or a *configuration* in Mesa.
759 A _crate_ is a unit of compilation and linking, as well as versioning,
760 distribution and runtime loading. A crate contains a _tree_ of nested
761 [module](#modules) scopes. The top level of this tree is a module that is
762 anonymous (from the point of view of paths within the module) and any item
763 within a crate has a canonical [module path](#paths) denoting its location
764 within the crate's module tree.
766 The Rust compiler is always invoked with a single source file as input, and
767 always produces a single output crate. The processing of that source file may
768 result in other source files being loaded as modules. Source files have the
771 A Rust source file describes a module, the name and location of which —
772 in the module tree of the current crate — are defined from outside the
773 source file: either by an explicit `mod_item` in a referencing source file, or
774 by the name of the crate itself.
776 Each source file contains a sequence of zero or more `item` definitions, and
777 may optionally begin with any number of `attributes` that apply to the
778 containing module. Attributes on the anonymous crate module define important
779 metadata that influences the behavior of the compiler.
782 # #![allow(unused_attribute)]
784 #![crate_name = "projx"]
786 // Specify the output type
787 #![crate_type = "lib"]
790 #![warn(non_camel_case_types)]
793 A crate that contains a `main` function can be compiled to an executable. If a
794 `main` function is present, its return type must be [`unit`](#primitive-types)
795 and it must take no arguments.
797 # Items and attributes
799 Crates contain [items](#items), each of which may have some number of
800 [attributes](#attributes) attached to it.
805 item : mod_item | fn_item | type_item | struct_item | enum_item
806 | static_item | trait_item | impl_item | extern_block ;
809 An _item_ is a component of a crate; some module items can be defined in crate
810 files, but most are defined in source files. Items are organized within a crate
811 by a nested set of [modules](#modules). Every crate has a single "outermost"
812 anonymous module; all further items within the crate have [paths](#paths)
813 within the module tree of the crate.
815 Items are entirely determined at compile-time, generally remain fixed during
816 execution, and may reside in read-only memory.
818 There are several kinds of item:
820 * [modules](#modules)
821 * [functions](#functions)
822 * [type definitions](#type-definitions)
823 * [structures](#structures)
824 * [enumerations](#enumerations)
825 * [static items](#static-items)
827 * [implementations](#implementations)
829 Some items form an implicit scope for the declaration of sub-items. In other
830 words, within a function or module, declarations of items can (in many cases)
831 be mixed with the statements, control blocks, and similar artifacts that
832 otherwise compose the item body. The meaning of these scoped items is the same
833 as if the item was declared outside the scope — it is still a static item
834 — except that the item's *path name* within the module namespace is
835 qualified by the name of the enclosing item, or is private to the enclosing
836 item (in the case of functions). The grammar specifies the exact locations in
837 which sub-item declarations may appear.
841 All items except modules may be *parameterized* by type. Type parameters are
842 given as a comma-separated list of identifiers enclosed in angle brackets
843 (`<...>`), after the name of the item and before its definition. The type
844 parameters of an item are considered "part of the name", not part of the type
845 of the item. A referencing [path](#paths) must (in principle) provide type
846 arguments as a list of comma-separated types enclosed within angle brackets, in
847 order to refer to the type-parameterized item. In practice, the type-inference
848 system can usually infer such argument types from context. There are no
849 general type-parametric types, only type-parametric items. That is, Rust has
850 no notion of type abstraction: there are no first-class "forall" types.
855 mod_item : "mod" ident ( ';' | '{' mod '}' );
856 mod : [ view_item | item ] * ;
859 A module is a container for zero or more [view items](#view-items) and zero or
860 more [items](#items). The view items manage the visibility of the items defined
861 within the module, as well as the visibility of names from outside the module
862 when referenced from inside the module.
864 A _module item_ is a module, surrounded in braces, named, and prefixed with the
865 keyword `mod`. A module item introduces a new, named module into the tree of
866 modules making up a crate. Modules can nest arbitrarily.
868 An example of a module:
872 type Complex = (f64, f64);
873 fn sin(f: f64) -> f64 {
877 fn cos(f: f64) -> f64 {
881 fn tan(f: f64) -> f64 {
888 Modules and types share the same namespace. Declaring a named type with
889 the same name as a module in scope is forbidden: that is, a type definition,
890 trait, struct, enumeration, or type parameter can't shadow the name of a module
891 in scope, or vice versa.
893 A module without a body is loaded from an external file, by default with the
894 same name as the module, plus the `.rs` extension. When a nested submodule is
895 loaded from an external file, it is loaded from a subdirectory path that
896 mirrors the module hierarchy.
899 // Load the `vec` module from `vec.rs`
903 // Load the `local_data` module from `thread/local_data.rs`
908 The directories and files used for loading external file modules can be
909 influenced with the `path` attribute.
912 #[path = "thread_files"]
914 // Load the `local_data` module from `thread_files/tls.rs`
923 view_item : extern_crate_decl | use_decl ;
926 A view item manages the namespace of a module. View items do not define new
927 items, but rather, simply change other items' visibility. There are two
930 * [`extern crate` declarations](#extern-crate-declarations)
931 * [`use` declarations](#use-declarations)
933 ##### Extern crate declarations
936 extern_crate_decl : "extern" "crate" crate_name
937 crate_name: ident | ( string_lit "as" ident )
940 An _`extern crate` declaration_ specifies a dependency on an external crate.
941 The external crate is then bound into the declaring scope as the `ident`
942 provided in the `extern_crate_decl`.
944 The external crate is resolved to a specific `soname` at compile time, and a
945 runtime linkage requirement to that `soname` is passed to the linker for
946 loading at runtime. The `soname` is resolved at compile time by scanning the
947 compiler's library path and matching the optional `crateid` provided as a
948 string literal against the `crateid` attributes that were declared on the
949 external crate when it was compiled. If no `crateid` is provided, a default
950 `name` attribute is assumed, equal to the `ident` given in the
953 Three examples of `extern crate` declarations:
958 extern crate std; // equivalent to: extern crate std as std;
960 extern crate "std" as ruststd; // linking to 'std' under another name
963 ##### Use declarations
966 use_decl : "pub" ? "use" [ path "as" ident
969 path_glob : ident [ "::" [ path_glob
971 | '{' path_item [ ',' path_item ] * '}' ;
973 path_item : ident | "mod" ;
976 A _use declaration_ creates one or more local name bindings synonymous with
977 some other [path](#paths). Usually a `use` declaration is used to shorten the
978 path required to refer to a module item. These declarations may appear at the
979 top of [modules](#modules) and [blocks](#blocks).
981 > **Note**: Unlike in many languages,
982 > `use` declarations in Rust do *not* declare linkage dependency with external crates.
983 > Rather, [`extern crate` declarations](#extern-crate-declarations) declare linkage dependencies.
985 Use declarations support a number of convenient shortcuts:
987 * Rebinding the target name as a new local name, using the syntax `use p::q::r as x;`.
988 * Simultaneously binding a list of paths differing only in their final element,
989 using the glob-like brace syntax `use a::b::{c,d,e,f};`
990 * Binding all paths matching a given prefix, using the asterisk wildcard syntax
992 * Simultaneously binding a list of paths differing only in their final element
993 and their immediate parent module, using the `mod` keyword, such as
994 `use a::b::{mod, c, d};`
996 An example of `use` declarations:
999 use std::iter::range_step;
1000 use std::option::Option::{Some, None};
1001 use std::collections::hash_map::{mod, HashMap};
1004 fn bar(map1: HashMap<String, uint>, map2: hash_map::HashMap<String, uint>){}
1007 // Equivalent to 'std::iter::range_step(0u, 10u, 2u);'
1008 range_step(0u, 10u, 2u);
1010 // Equivalent to 'foo(vec![std::option::Option::Some(1.0f64),
1011 // std::option::Option::None]);'
1012 foo(vec![Some(1.0f64), None]);
1014 // Both `hash_map` and `HashMap` are in scope.
1015 let map1 = HashMap::new();
1016 let map2 = hash_map::HashMap::new();
1021 Like items, `use` declarations are private to the containing module, by
1022 default. Also like items, a `use` declaration can be public, if qualified by
1023 the `pub` keyword. Such a `use` declaration serves to _re-export_ a name. A
1024 public `use` declaration can therefore _redirect_ some public name to a
1025 different target definition: even a definition with a private canonical path,
1026 inside a different module. If a sequence of such redirections form a cycle or
1027 cannot be resolved unambiguously, they represent a compile-time error.
1029 An example of re-exporting:
1034 pub use quux::foo::{bar, baz};
1043 In this example, the module `quux` re-exports two public names defined in
1046 Also note that the paths contained in `use` items are relative to the crate
1047 root. So, in the previous example, the `use` refers to `quux::foo::{bar,
1048 baz}`, and not simply to `foo::{bar, baz}`. This also means that top-level
1049 module declarations should be at the crate root if direct usage of the declared
1050 modules within `use` items is desired. It is also possible to use `self` and
1051 `super` at the beginning of a `use` item to refer to the current and direct
1052 parent modules respectively. All rules regarding accessing declared modules in
1053 `use` declarations applies to both module declarations and `extern crate`
1056 An example of what will and will not work for `use` items:
1059 # #![allow(unused_imports)]
1060 use foo::core::iter; // good: foo is at the root of the crate
1061 use foo::baz::foobaz; // good: foo is at the root of the crate
1066 use foo::core::iter; // good: foo is at crate root
1067 // use core::iter; // bad: native is not at the crate root
1068 use self::baz::foobaz; // good: self refers to module 'foo'
1069 use foo::bar::foobar; // good: foo is at crate root
1076 use super::bar::foobar; // good: super refers to module 'foo'
1086 A _function item_ defines a sequence of [statements](#statements) and an
1087 optional final [expression](#expressions), along with a name and a set of
1088 parameters. Functions are declared with the keyword `fn`. Functions declare a
1089 set of *input* [*slots*](#memory-slots) as parameters, through which the caller
1090 passes arguments into the function, and an *output* [*slot*](#memory-slots)
1091 through which the function passes results back to the caller.
1093 A function may also be copied into a first class *value*, in which case the
1094 value has the corresponding [*function type*](#function-types), and can be used
1095 otherwise exactly as a function item (with a minor additional cost of calling
1096 the function indirectly).
1098 Every control path in a function logically ends with a `return` expression or a
1099 diverging expression. If the outermost block of a function has a
1100 value-producing expression in its final-expression position, that expression is
1101 interpreted as an implicit `return` expression applied to the final-expression.
1103 An example of a function:
1106 fn add(x: int, y: int) -> int {
1111 As with `let` bindings, function arguments are irrefutable patterns, so any
1112 pattern that is valid in a let binding is also valid as an argument.
1115 fn first((value, _): (int, int)) -> int { value }
1119 #### Generic functions
1121 A _generic function_ allows one or more _parameterized types_ to appear in its
1122 signature. Each type parameter must be explicitly declared, in an
1123 angle-bracket-enclosed, comma-separated list following the function name.
1126 fn iter<T>(seq: &[T], f: |T|) {
1127 for elt in seq.iter() { f(elt); }
1129 fn map<T, U>(seq: &[T], f: |T| -> U) -> Vec<U> {
1130 let mut acc = vec![];
1131 for elt in seq.iter() { acc.push(f(elt)); }
1136 Inside the function signature and body, the name of the type parameter can be
1137 used as a type name.
1139 When a generic function is referenced, its type is instantiated based on the
1140 context of the reference. For example, calling the `iter` function defined
1141 above on `[1, 2]` will instantiate type parameter `T` with `int`, and require
1142 the closure parameter to have type `fn(int)`.
1144 The type parameters can also be explicitly supplied in a trailing
1145 [path](#paths) component after the function name. This might be necessary if
1146 there is not sufficient context to determine the type parameters. For example,
1147 `mem::size_of::<u32>() == 4`.
1149 Since a parameter type is opaque to the generic function, the set of operations
1150 that can be performed on it is limited. Values of parameter type can only be
1154 fn id<T>(x: T) -> T { x }
1157 Similarly, [trait](#traits) bounds can be specified for type parameters to
1158 allow methods with that trait to be called on values of that type.
1162 Unsafe operations are those that potentially violate the memory-safety
1163 guarantees of Rust's static semantics.
1165 The following language level features cannot be used in the safe subset of
1168 - Dereferencing a [raw pointer](#pointer-types).
1169 - Reading or writing a [mutable static variable](#mutable-statics).
1170 - Calling an unsafe function (including an intrinsic or foreign function).
1172 ##### Unsafe functions
1174 Unsafe functions are functions that are not safe in all contexts and/or for all
1175 possible inputs. Such a function must be prefixed with the keyword `unsafe` and
1176 can only be called from an `unsafe` block or another `unsafe` function.
1180 A block of code can be prefixed with the `unsafe` keyword, to permit calling
1181 `unsafe` functions or dereferencing raw pointers within a safe function.
1183 When a programmer has sufficient conviction that a sequence of potentially
1184 unsafe operations is actually safe, they can encapsulate that sequence (taken
1185 as a whole) within an `unsafe` block. The compiler will consider uses of such
1186 code safe, in the surrounding context.
1188 Unsafe blocks are used to wrap foreign libraries, make direct use of hardware
1189 or implement features not directly present in the language. For example, Rust
1190 provides the language features necessary to implement memory-safe concurrency
1191 in the language but the implementation of threads and message passing is in the
1194 Rust's type system is a conservative approximation of the dynamic safety
1195 requirements, so in some cases there is a performance cost to using safe code.
1196 For example, a doubly-linked list is not a tree structure and can only be
1197 represented with reference-counted pointers in safe code. By using `unsafe`
1198 blocks to represent the reverse links as raw pointers, it can be implemented
1201 ##### Behavior considered undefined
1203 The following is a list of behavior which is forbidden in all Rust code,
1204 including within `unsafe` blocks and `unsafe` functions. Type checking provides
1205 the guarantee that these issues are never caused by safe code.
1208 * Dereferencing a null/dangling raw pointer
1209 * Mutating an immutable value/reference without `UnsafeCell`
1210 * Reads of [undef](http://llvm.org/docs/LangRef.html#undefined-values)
1211 (uninitialized) memory
1212 * Breaking the [pointer aliasing
1213 rules](http://llvm.org/docs/LangRef.html#pointer-aliasing-rules)
1214 with raw pointers (a subset of the rules used by C)
1215 * Invoking undefined behavior via compiler intrinsics:
1216 * Indexing outside of the bounds of an object with `std::ptr::offset`
1217 (`offset` intrinsic), with
1218 the exception of one byte past the end which is permitted.
1219 * Using `std::ptr::copy_nonoverlapping_memory` (`memcpy32`/`memcpy64`
1220 instrinsics) on overlapping buffers
1221 * Invalid values in primitive types, even in private fields/locals:
1222 * Dangling/null references or boxes
1223 * A value other than `false` (0) or `true` (1) in a `bool`
1224 * A discriminant in an `enum` not included in the type definition
1225 * A value in a `char` which is a surrogate or above `char::MAX`
1226 * non-UTF-8 byte sequences in a `str`
1227 * Unwinding into Rust from foreign code or unwinding from Rust into foreign
1228 code. Rust's failure system is not compatible with exception handling in
1229 other languages. Unwinding must be caught and handled at FFI boundaries.
1231 ##### Behaviour not considered unsafe
1233 This is a list of behaviour not considered *unsafe* in Rust terms, but that may
1237 * Reading data from private fields (`std::repr`)
1238 * Leaks due to reference count cycles, even in the global heap
1239 * Exiting without calling destructors
1241 * Accessing/modifying the file system
1242 * Unsigned integer overflow (well-defined as wrapping)
1243 * Signed integer overflow (well-defined as two's complement representation
1246 #### Diverging functions
1248 A special kind of function can be declared with a `!` character where the
1249 output slot type would normally be. For example:
1252 fn my_err(s: &str) -> ! {
1258 We call such functions "diverging" because they never return a value to the
1259 caller. Every control path in a diverging function must end with a `panic!()` or
1260 a call to another diverging function on every control path. The `!` annotation
1261 does *not* denote a type. Rather, the result type of a diverging function is a
1262 special type called $\bot$ ("bottom") that unifies with any type. Rust has no
1265 It might be necessary to declare a diverging function because as mentioned
1266 previously, the typechecker checks that every control path in a function ends
1267 with a [`return`](#return-expressions) or diverging expression. So, if `my_err`
1268 were declared without the `!` annotation, the following code would not
1272 # fn my_err(s: &str) -> ! { panic!() }
1274 fn f(i: int) -> int {
1279 my_err("Bad number!");
1284 This will not compile without the `!` annotation on `my_err`, since the `else`
1285 branch of the conditional in `f` does not return an `int`, as required by the
1286 signature of `f`. Adding the `!` annotation to `my_err` informs the
1287 typechecker that, should control ever enter `my_err`, no further type judgments
1288 about `f` need to hold, since control will never resume in any context that
1289 relies on those judgments. Thus the return type on `f` only needs to reflect
1290 the `if` branch of the conditional.
1292 #### Extern functions
1294 Extern functions are part of Rust's foreign function interface, providing the
1295 opposite functionality to [external blocks](#external-blocks). Whereas
1296 external blocks allow Rust code to call foreign code, extern functions with
1297 bodies defined in Rust code _can be called by foreign code_. They are defined
1298 in the same way as any other Rust function, except that they have the `extern`
1302 // Declares an extern fn, the ABI defaults to "C"
1303 extern fn new_int() -> int { 0 }
1305 // Declares an extern fn with "stdcall" ABI
1306 extern "stdcall" fn new_int_stdcall() -> int { 0 }
1309 Unlike normal functions, extern fns have an `extern "ABI" fn()`. This is the
1310 same type as the functions declared in an extern block.
1313 # extern fn new_int() -> int { 0 }
1314 let fptr: extern "C" fn() -> int = new_int;
1317 Extern functions may be called directly from Rust code as Rust uses large,
1318 contiguous stack segments like C.
1322 A _type alias_ defines a new name for an existing [type](#types). Type
1323 aliases are declared with the keyword `type`. Every value has a single,
1324 specific type; the type-specified aspects of a value include:
1326 * Whether the value is composed of sub-values or is indivisible.
1327 * Whether the value represents textual or numerical information.
1328 * Whether the value represents integral or floating-point information.
1329 * The sequence of memory operations required to access the value.
1330 * The [kind](#type-kinds) of the type.
1332 For example, the type `(u8, u8)` defines the set of immutable values that are
1333 composite pairs, each containing two unsigned 8-bit integers accessed by
1334 pattern-matching and laid out in memory with the `x` component preceding the
1338 type Point = (u8, u8);
1339 let p: Point = (41, 68);
1344 A _structure_ is a nominal [structure type](#structure-types) defined with the
1347 An example of a `struct` item and its use:
1350 struct Point {x: int, y: int}
1351 let p = Point {x: 10, y: 11};
1355 A _tuple structure_ is a nominal [tuple type](#tuple-types), also defined with
1356 the keyword `struct`. For example:
1359 struct Point(int, int);
1360 let p = Point(10, 11);
1361 let px: int = match p { Point(x, _) => x };
1364 A _unit-like struct_ is a structure without any fields, defined by leaving off
1365 the list of fields entirely. Such types will have a single value, just like
1366 the [unit value `()`](#unit-and-boolean-literals) of the unit type. For
1371 let c = [Cookie, Cookie, Cookie, Cookie];
1374 The precise memory layout of a structure is not specified. One can specify a
1375 particular layout using the [`repr` attribute](#ffi-attributes).
1379 An _enumeration_ is a simultaneous definition of a nominal [enumerated
1380 type](#enumerated-types) as well as a set of *constructors*, that can be used
1381 to create or pattern-match values of the corresponding enumerated type.
1383 Enumerations are declared with the keyword `enum`.
1385 An example of an `enum` item and its use:
1393 let mut a: Animal = Animal::Dog;
1397 Enumeration constructors can have either named or unnamed fields:
1400 # #![feature(struct_variant)]
1404 Cat { name: String, weight: f64 }
1407 let mut a: Animal = Animal::Dog("Cocoa".to_string(), 37.2);
1408 a = Animal::Cat { name: "Spotty".to_string(), weight: 2.7 };
1412 In this example, `Cat` is a _struct-like enum variant_,
1413 whereas `Dog` is simply called an enum variant.
1418 const_item : "const" ident ':' type '=' expr ';' ;
1421 A *constant item* is a named _constant value_ which is not associated with a
1422 specific memory location in the program. Constants are essentially inlined
1423 wherever they are used, meaning that they are copied directly into the relevant
1424 context when used. References to the same constant are not necessarily
1425 guaranteed to refer to the same memory address.
1427 Constant values must not have destructors, and otherwise permit most forms of
1428 data. Constants may refer to the address of other constants, in which case the
1429 address will have the `static` lifetime. The compiler is, however, still at
1430 liberty to translate the constant many times, so the address referred to may not
1433 Constants must be explicitly typed. The type may be `bool`, `char`, a number, or
1434 a type derived from those primitive types. The derived types are references with
1435 the `static` lifetime, fixed-size arrays, tuples, enum variants, and structs.
1438 const BIT1: uint = 1 << 0;
1439 const BIT2: uint = 1 << 1;
1441 const BITS: [uint; 2] = [BIT1, BIT2];
1442 const STRING: &'static str = "bitstring";
1444 struct BitsNStrings<'a> {
1449 const BITS_N_STRINGS: BitsNStrings<'static> = BitsNStrings {
1458 static_item : "static" ident ':' type '=' expr ';' ;
1461 A *static item* is similar to a *constant*, except that it represents a precise
1462 memory location in the program. A static is never "inlined" at the usage site,
1463 and all references to it refer to the same memory location. Static items have
1464 the `static` lifetime, which outlives all other lifetimes in a Rust program.
1465 Static items may be placed in read-only memory if they do not contain any
1466 interior mutability.
1468 Statics may contain interior mutability through the `UnsafeCell` language item.
1469 All access to a static is safe, but there are a number of restrictions on
1472 * Statics may not contain any destructors.
1473 * The types of static values must ascribe to `Sync` to allow threadsafe access.
1474 * Statics may not refer to other statics by value, only by reference.
1475 * Constants cannot refer to statics.
1477 Constants should in general be preferred over statics, unless large amounts of
1478 data are being stored, or single-address and mutability properties are required.
1481 use std::sync::atomic;
1483 // Note that ATOMIC_UINT_INIT is a *const*, but it may be used to initialize a
1484 // static. This static can be modified, so it is not placed in read-only memory.
1485 static COUNTER: atomic::AtomicUint = atomic::ATOMIC_UINT_INIT;
1487 // This table is a candidate to be placed in read-only memory.
1488 static TABLE: &'static [uint] = &[1, 2, 3, /* ... */];
1490 for slot in TABLE.iter() {
1491 println!("{}", slot);
1493 COUNTER.fetch_add(1, atomic::SeqCst);
1496 #### Mutable statics
1498 If a static item is declared with the `mut` keyword, then it is allowed to
1499 be modified by the program. One of Rust's goals is to make concurrency bugs
1500 hard to run into, and this is obviously a very large source of race conditions
1501 or other bugs. For this reason, an `unsafe` block is required when either
1502 reading or writing a mutable static variable. Care should be taken to ensure
1503 that modifications to a mutable static are safe with respect to other threads
1504 running in the same process.
1506 Mutable statics are still very useful, however. They can be used with C
1507 libraries and can also be bound from C libraries (in an `extern` block).
1510 # fn atomic_add(_: &mut uint, _: uint) -> uint { 2 }
1512 static mut LEVELS: uint = 0;
1514 // This violates the idea of no shared state, and this doesn't internally
1515 // protect against races, so this function is `unsafe`
1516 unsafe fn bump_levels_unsafe1() -> uint {
1522 // Assuming that we have an atomic_add function which returns the old value,
1523 // this function is "safe" but the meaning of the return value may not be what
1524 // callers expect, so it's still marked as `unsafe`
1525 unsafe fn bump_levels_unsafe2() -> uint {
1526 return atomic_add(&mut LEVELS, 1);
1530 Mutable statics have the same restrictions as normal statics, except that the
1531 type of the value is not required to ascribe to `Sync`.
1535 A _trait_ describes a set of method types.
1537 Traits can include default implementations of methods, written in terms of some
1538 unknown [`self` type](#self-types); the `self` type may either be completely
1539 unspecified, or constrained by some other trait.
1541 Traits are implemented for specific types through separate
1542 [implementations](#implementations).
1545 # type Surface = int;
1546 # type BoundingBox = int;
1548 fn draw(&self, Surface);
1549 fn bounding_box(&self) -> BoundingBox;
1553 This defines a trait with two methods. All values that have
1554 [implementations](#implementations) of this trait in scope can have their
1555 `draw` and `bounding_box` methods called, using `value.bounding_box()`
1556 [syntax](#method-call-expressions).
1558 Type parameters can be specified for a trait to make it generic. These appear
1559 after the trait name, using the same syntax used in [generic
1560 functions](#generic-functions).
1564 fn len(&self) -> uint;
1565 fn elt_at(&self, n: uint) -> T;
1566 fn iter(&self, |T|);
1570 Generic functions may use traits as _bounds_ on their type parameters. This
1571 will have two effects: only types that have the trait may instantiate the
1572 parameter, and within the generic function, the methods of the trait can be
1573 called on values that have the parameter's type. For example:
1576 # type Surface = int;
1577 # trait Shape { fn draw(&self, Surface); }
1578 fn draw_twice<T: Shape>(surface: Surface, sh: T) {
1584 Traits also define an [object type](#object-types) with the same name as the
1585 trait. Values of this type are created by [casting](#type-cast-expressions)
1586 pointer values (pointing to a type for which an implementation of the given
1587 trait is in scope) to pointers to the trait name, used as a type.
1591 # impl Shape for int { }
1592 # let mycircle = 0i;
1593 let myshape: Box<Shape> = box mycircle as Box<Shape>;
1596 The resulting value is a box containing the value that was cast, along with
1597 information that identifies the methods of the implementation that was used.
1598 Values with a trait type can have [methods called](#method-call-expressions) on
1599 them, for any method in the trait, and can be used to instantiate type
1600 parameters that are bounded by the trait.
1602 Trait methods may be static, which means that they lack a `self` argument.
1603 This means that they can only be called with function call syntax (`f(x)`) and
1604 not method call syntax (`obj.f()`). The way to refer to the name of a static
1605 method is to qualify it with the trait name, treating the trait name like a
1606 module. For example:
1610 fn from_int(n: int) -> Self;
1613 fn from_int(n: int) -> f64 { n as f64 }
1615 let x: f64 = Num::from_int(42);
1618 Traits may inherit from other traits. For example, in
1621 trait Shape { fn area() -> f64; }
1622 trait Circle : Shape { fn radius() -> f64; }
1625 the syntax `Circle : Shape` means that types that implement `Circle` must also
1626 have an implementation for `Shape`. Multiple supertraits are separated by `+`,
1627 `trait Circle : Shape + PartialEq { }`. In an implementation of `Circle` for a
1628 given type `T`, methods can refer to `Shape` methods, since the typechecker
1629 checks that any type with an implementation of `Circle` also has an
1630 implementation of `Shape`.
1632 In type-parameterized functions, methods of the supertrait may be called on
1633 values of subtrait-bound type parameters. Referring to the previous example of
1634 `trait Circle : Shape`:
1637 # trait Shape { fn area(&self) -> f64; }
1638 # trait Circle : Shape { fn radius(&self) -> f64; }
1639 fn radius_times_area<T: Circle>(c: T) -> f64 {
1640 // `c` is both a Circle and a Shape
1641 c.radius() * c.area()
1645 Likewise, supertrait methods may also be called on trait objects.
1648 # trait Shape { fn area(&self) -> f64; }
1649 # trait Circle : Shape { fn radius(&self) -> f64; }
1650 # impl Shape for int { fn area(&self) -> f64 { 0.0 } }
1651 # impl Circle for int { fn radius(&self) -> f64 { 0.0 } }
1653 let mycircle = box mycircle as Box<Circle>;
1654 let nonsense = mycircle.radius() * mycircle.area();
1659 An _implementation_ is an item that implements a [trait](#traits) for a
1662 Implementations are defined with the keyword `impl`.
1665 # struct Point {x: f64, y: f64};
1666 # impl Copy for Point {}
1667 # type Surface = int;
1668 # struct BoundingBox {x: f64, y: f64, width: f64, height: f64};
1669 # trait Shape { fn draw(&self, Surface); fn bounding_box(&self) -> BoundingBox; }
1670 # fn do_draw_circle(s: Surface, c: Circle) { }
1676 impl Copy for Circle {}
1678 impl Shape for Circle {
1679 fn draw(&self, s: Surface) { do_draw_circle(s, *self); }
1680 fn bounding_box(&self) -> BoundingBox {
1681 let r = self.radius;
1682 BoundingBox{x: self.center.x - r, y: self.center.y - r,
1683 width: 2.0 * r, height: 2.0 * r}
1688 It is possible to define an implementation without referring to a trait. The
1689 methods in such an implementation can only be used as direct calls on the
1690 values of the type that the implementation targets. In such an implementation,
1691 the trait type and `for` after `impl` are omitted. Such implementations are
1692 limited to nominal types (enums, structs), and the implementation must appear
1693 in the same module or a sub-module as the `self` type:
1696 struct Point {x: int, y: int}
1700 println!("Point is at ({}, {})", self.x, self.y);
1704 let my_point = Point {x: 10, y:11};
1708 When a trait _is_ specified in an `impl`, all methods declared as part of the
1709 trait must be implemented, with matching types and type parameter counts.
1711 An implementation can take type parameters, which can be different from the
1712 type parameters taken by the trait it implements. Implementation parameters
1713 are written after the `impl` keyword.
1717 impl<T> Seq<T> for Vec<T> {
1720 impl Seq<bool> for u32 {
1721 /* Treat the integer as a sequence of bits */
1728 extern_block_item : "extern" '{' extern_block '}' ;
1729 extern_block : [ foreign_fn ] * ;
1732 External blocks form the basis for Rust's foreign function interface.
1733 Declarations in an external block describe symbols in external, non-Rust
1736 Functions within external blocks are declared in the same way as other Rust
1737 functions, with the exception that they may not have a body and are instead
1738 terminated by a semicolon.
1742 use libc::{c_char, FILE};
1745 fn fopen(filename: *const c_char, mode: *const c_char) -> *mut FILE;
1750 Functions within external blocks may be called by Rust code, just like
1751 functions defined in Rust. The Rust compiler automatically translates between
1752 the Rust ABI and the foreign ABI.
1754 A number of [attributes](#attributes) control the behavior of external blocks.
1756 By default external blocks assume that the library they are calling uses the
1757 standard C "cdecl" ABI. Other ABIs may be specified using an `abi` string, as
1761 // Interface to the Windows API
1762 extern "stdcall" { }
1765 The `link` attribute allows the name of the library to be specified. When
1766 specified the compiler will attempt to link against the native library of the
1770 #[link(name = "crypto")]
1774 The type of a function declared in an extern block is `extern "abi" fn(A1, ...,
1775 An) -> R`, where `A1...An` are the declared types of its arguments and `R` is
1776 the declared return type.
1778 ## Visibility and Privacy
1780 These two terms are often used interchangeably, and what they are attempting to
1781 convey is the answer to the question "Can this item be used at this location?"
1783 Rust's name resolution operates on a global hierarchy of namespaces. Each level
1784 in the hierarchy can be thought of as some item. The items are one of those
1785 mentioned above, but also include external crates. Declaring or defining a new
1786 module can be thought of as inserting a new tree into the hierarchy at the
1787 location of the definition.
1789 To control whether interfaces can be used across modules, Rust checks each use
1790 of an item to see whether it should be allowed or not. This is where privacy
1791 warnings are generated, or otherwise "you used a private item of another module
1792 and weren't allowed to."
1794 By default, everything in Rust is *private*, with one exception. Enum variants
1795 in a `pub` enum are also public by default. You are allowed to alter this
1796 default visibility with the `priv` keyword. When an item is declared as `pub`,
1797 it can be thought of as being accessible to the outside world. For example:
1800 # #![allow(missing_copy_implementations)]
1802 // Declare a private struct
1805 // Declare a public struct with a private field
1810 // Declare a public enum with two public variants
1812 PubliclyAccessibleState,
1813 PubliclyAccessibleState2,
1817 With the notion of an item being either public or private, Rust allows item
1818 accesses in two cases:
1820 1. If an item is public, then it can be used externally through any of its
1822 2. If an item is private, it may be accessed by the current module and its
1825 These two cases are surprisingly powerful for creating module hierarchies
1826 exposing public APIs while hiding internal implementation details. To help
1827 explain, here's a few use cases and what they would entail.
1829 * A library developer needs to expose functionality to crates which link
1830 against their library. As a consequence of the first case, this means that
1831 anything which is usable externally must be `pub` from the root down to the
1832 destination item. Any private item in the chain will disallow external
1835 * A crate needs a global available "helper module" to itself, but it doesn't
1836 want to expose the helper module as a public API. To accomplish this, the
1837 root of the crate's hierarchy would have a private module which then
1838 internally has a "public api". Because the entire crate is a descendant of
1839 the root, then the entire local crate can access this private module through
1842 * When writing unit tests for a module, it's often a common idiom to have an
1843 immediate child of the module to-be-tested named `mod test`. This module
1844 could access any items of the parent module through the second case, meaning
1845 that internal implementation details could also be seamlessly tested from the
1848 In the second case, it mentions that a private item "can be accessed" by the
1849 current module and its descendants, but the exact meaning of accessing an item
1850 depends on what the item is. Accessing a module, for example, would mean
1851 looking inside of it (to import more items). On the other hand, accessing a
1852 function would mean that it is invoked. Additionally, path expressions and
1853 import statements are considered to access an item in the sense that the
1854 import/expression is only valid if the destination is in the current visibility
1857 Here's an example of a program which exemplifies the three cases outlined
1861 // This module is private, meaning that no external crate can access this
1862 // module. Because it is private at the root of this current crate, however, any
1863 // module in the crate may access any publicly visible item in this module.
1864 mod crate_helper_module {
1866 // This function can be used by anything in the current crate
1867 pub fn crate_helper() {}
1869 // This function *cannot* be used by anything else in the crate. It is not
1870 // publicly visible outside of the `crate_helper_module`, so only this
1871 // current module and its descendants may access it.
1872 fn implementation_detail() {}
1875 // This function is "public to the root" meaning that it's available to external
1876 // crates linking against this one.
1877 pub fn public_api() {}
1879 // Similarly to 'public_api', this module is public so external crates may look
1882 use crate_helper_module;
1884 pub fn my_method() {
1885 // Any item in the local crate may invoke the helper module's public
1886 // interface through a combination of the two rules above.
1887 crate_helper_module::crate_helper();
1890 // This function is hidden to any module which is not a descendant of
1892 fn my_implementation() {}
1898 fn test_my_implementation() {
1899 // Because this module is a descendant of `submodule`, it's allowed
1900 // to access private items inside of `submodule` without a privacy
1902 super::my_implementation();
1910 For a rust program to pass the privacy checking pass, all paths must be valid
1911 accesses given the two rules above. This includes all use statements,
1912 expressions, types, etc.
1914 ### Re-exporting and Visibility
1916 Rust allows publicly re-exporting items through a `pub use` directive. Because
1917 this is a public directive, this allows the item to be used in the current
1918 module through the rules above. It essentially allows public access into the
1919 re-exported item. For example, this program is valid:
1922 pub use self::implementation as api;
1924 mod implementation {
1931 This means that any external crate referencing `implementation::f` would
1932 receive a privacy violation, while the path `api::f` would be allowed.
1934 When re-exporting a private item, it can be thought of as allowing the "privacy
1935 chain" being short-circuited through the reexport instead of passing through
1936 the namespace hierarchy as it normally would.
1941 attribute : "#!" ? '[' meta_item ']' ;
1942 meta_item : ident [ '=' literal
1943 | '(' meta_seq ')' ] ? ;
1944 meta_seq : meta_item [ ',' meta_seq ] ? ;
1947 Any item declaration may have an _attribute_ applied to it. Attributes in Rust
1948 are modeled on Attributes in ECMA-335, with the syntax coming from ECMA-334
1949 (C#). An attribute is a general, free-form metadatum that is interpreted
1950 according to name, convention, and language and compiler version. Attributes
1951 may appear as any of:
1953 * A single identifier, the attribute name
1954 * An identifier followed by the equals sign '=' and a literal, providing a
1956 * An identifier followed by a parenthesized list of sub-attribute arguments
1958 Attributes with a bang ("!") after the hash ("#") apply to the item that the
1959 attribute is declared within. Attributes that do not have a bang after the hash
1960 apply to the item that follows the attribute.
1962 An example of attributes:
1965 // General metadata applied to the enclosing module or crate.
1966 #![crate_type = "lib"]
1968 // A function marked as a unit test
1974 // A conditionally-compiled module
1975 #[cfg(target_os="linux")]
1980 // A lint attribute used to suppress a warning/error
1981 #[allow(non_camel_case_types)]
1985 > **Note:** At some point in the future, the compiler will distinguish between
1986 > language-reserved and user-available attributes. Until then, there is
1987 > effectively no difference between an attribute handled by a loadable syntax
1988 > extension and the compiler.
1990 ### Crate-only attributes
1992 - `crate_name` - specify the this crate's crate name.
1993 - `crate_type` - see [linkage](#linkage).
1994 - `feature` - see [compiler features](#compiler-features).
1995 - `no_builtins` - disable optimizing certain code patterns to invocations of
1996 library functions that are assumed to exist
1997 - `no_main` - disable emitting the `main` symbol. Useful when some other
1998 object being linked to defines `main`.
1999 - `no_start` - disable linking to the `native` crate, which specifies the
2000 "start" language item.
2001 - `no_std` - disable linking to the `std` crate.
2003 ### Module-only attributes
2005 - `macro_escape` - macros defined in this module will be visible in the
2006 module's parent, after this module has been included.
2007 - `no_implicit_prelude` - disable injecting `use std::prelude::*` in this
2009 - `path` - specifies the file to load the module from. `#[path="foo.rs"] mod
2010 bar;` is equivalent to `mod bar { /* contents of foo.rs */ }`. The path is
2011 taken relative to the directory that the current module is in.
2013 ### Function-only attributes
2015 - `main` - indicates that this function should be passed to the entry point,
2016 rather than the function in the crate root named `main`.
2017 - `plugin_registrar` - mark this function as the registration point for
2018 [compiler plugins][plugin], such as loadable syntax extensions.
2019 - `start` - indicates that this function should be used as the entry point,
2020 overriding the "start" language item. See the "start" [language
2021 item](#language-items) for more details.
2022 - `test` - indicates that this function is a test function, to only be compiled
2023 in case of `--test`.
2025 ### Static-only attributes
2027 - `thread_local` - on a `static mut`, this signals that the value of this
2028 static may change depending on the current thread. The exact consequences of
2029 this are implementation-defined.
2033 On an `extern` block, the following attributes are interpreted:
2035 - `link_args` - specify arguments to the linker, rather than just the library
2036 name and type. This is feature gated and the exact behavior is
2037 implementation-defined (due to variety of linker invocation syntax).
2038 - `link` - indicate that a native library should be linked to for the
2039 declarations in this block to be linked correctly. `link` supports an optional `kind`
2040 key with three possible values: `dylib`, `static`, and `framework`. See [external blocks](#external-blocks) for more about external blocks. Two
2041 examples: `#[link(name = "readline")]` and
2042 `#[link(name = "CoreFoundation", kind = "framework")]`.
2044 On declarations inside an `extern` block, the following attributes are
2047 - `link_name` - the name of the symbol that this function or static should be
2049 - `linkage` - on a static, this specifies the [linkage
2050 type](http://llvm.org/docs/LangRef.html#linkage-types).
2054 - `repr` - on C-like enums, this sets the underlying type used for
2055 representation. Takes one argument, which is the primitive
2056 type this enum should be represented for, or `C`, which specifies that it
2057 should be the default `enum` size of the C ABI for that platform. Note that
2058 enum representation in C is undefined, and this may be incorrect when the C
2059 code is compiled with certain flags.
2063 - `repr` - specifies the representation to use for this struct. Takes a list
2064 of options. The currently accepted ones are `C` and `packed`, which may be
2065 combined. `C` will use a C ABI compatible struct layout, and `packed` will
2066 remove any padding between fields (note that this is very fragile and may
2067 break platforms which require aligned access).
2069 ### Miscellaneous attributes
2071 - `export_name` - on statics and functions, this determines the name of the
2073 - `link_section` - on statics and functions, this specifies the section of the
2074 object file that this item's contents will be placed into.
2075 - `macro_export` - export a macro for cross-crate usage.
2076 - `no_mangle` - on any item, do not apply the standard name mangling. Set the
2077 symbol for this item to its identifier.
2078 - `packed` - on structs or enums, eliminate any padding that would be used to
2080 - `phase` - on `extern crate` statements, allows specifying which "phase" of
2081 compilation the crate should be loaded for. Currently, there are two
2082 choices: `link` and `plugin`. `link` is the default. `plugin` will [load the
2083 crate at compile-time][plugin] and use any syntax extensions or lints that the crate
2084 defines. They can both be specified, `#[phase(link, plugin)]` to use a crate
2085 both at runtime and compiletime.
2086 - `simd` - on certain tuple structs, derive the arithmetic operators, which
2087 lower to the target's SIMD instructions, if any; the `simd` feature gate
2088 is necessary to use this attribute.
2089 - `static_assert` - on statics whose type is `bool`, terminates compilation
2090 with an error if it is not initialized to `true`.
2091 - `unsafe_destructor` - allow implementations of the "drop" language item
2092 where the type it is implemented for does not implement the "send" language
2093 item; the `unsafe_destructor` feature gate is needed to use this attribute
2094 - `unsafe_no_drop_flag` - on structs, remove the flag that prevents
2095 destructors from being run twice. Destructors might be run multiple times on
2096 the same object with this attribute.
2097 - `doc` - Doc comments such as `/// foo` are equivalent to `#[doc = "foo"]`.
2099 ### Conditional compilation
2101 Sometimes one wants to have different compiler outputs from the same code,
2102 depending on build target, such as targeted operating system, or to enable
2105 There are two kinds of configuration options, one that is either defined or not
2106 (`#[cfg(foo)]`), and the other that contains a string that can be checked
2107 against (`#[cfg(bar = "baz")]` (currently only compiler-defined configuration
2108 options can have the latter form).
2111 // The function is only included in the build when compiling for OSX
2112 #[cfg(target_os = "macos")]
2117 // This function is only included when either foo or bar is defined
2118 #[cfg(any(foo, bar))]
2119 fn needs_foo_or_bar() {
2123 // This function is only included when compiling for a unixish OS with a 32-bit
2125 #[cfg(all(unix, target_word_size = "32"))]
2126 fn on_32bit_unix() {
2130 // This function is only included when foo is not defined
2132 fn needs_not_foo() {
2137 This illustrates some conditional compilation can be achieved using the
2138 `#[cfg(...)]` attribute. `any`, `all` and `not` can be used to assemble
2139 arbitrarily complex configurations through nesting.
2141 The following configurations must be defined by the implementation:
2143 * `target_arch = "..."`. Target CPU architecture, such as `"x86"`, `"x86_64"`
2144 `"mips"`, `"arm"`, or `"aarch64"`.
2145 * `target_endian = "..."`. Endianness of the target CPU, either `"little"` or
2147 * `target_family = "..."`. Operating system family of the target, e. g.
2148 `"unix"` or `"windows"`. The value of this configuration option is defined
2149 as a configuration itself, like `unix` or `windows`.
2150 * `target_os = "..."`. Operating system of the target, examples include
2151 `"win32"`, `"macos"`, `"linux"`, `"android"`, `"freebsd"` or `"dragonfly"`.
2152 * `target_word_size = "..."`. Target word size in bits. This is set to `"32"`
2153 for targets with 32-bit pointers, and likewise set to `"64"` for 64-bit
2155 * `unix`. See `target_family`.
2156 * `windows`. See `target_family`.
2158 ### Lint check attributes
2160 A lint check names a potentially undesirable coding pattern, such as
2161 unreachable code or omitted documentation, for the static entity to which the
2164 For any lint check `C`:
2166 * `allow(C)` overrides the check for `C` so that violations will go
2168 * `deny(C)` signals an error after encountering a violation of `C`,
2169 * `forbid(C)` is the same as `deny(C)`, but also forbids changing the lint
2171 * `warn(C)` warns about violations of `C` but continues compilation.
2173 The lint checks supported by the compiler can be found via `rustc -W help`,
2174 along with their default settings. [Compiler
2175 plugins](guide-plugin.html#lint-plugins) can provide additional lint checks.
2179 // Missing documentation is ignored here
2180 #[allow(missing_docs)]
2181 pub fn undocumented_one() -> int { 1 }
2183 // Missing documentation signals a warning here
2184 #[warn(missing_docs)]
2185 pub fn undocumented_too() -> int { 2 }
2187 // Missing documentation signals an error here
2188 #[deny(missing_docs)]
2189 pub fn undocumented_end() -> int { 3 }
2193 This example shows how one can use `allow` and `warn` to toggle a particular
2197 #[warn(missing_docs)]
2199 #[allow(missing_docs)]
2201 // Missing documentation is ignored here
2202 pub fn undocumented_one() -> int { 1 }
2204 // Missing documentation signals a warning here,
2205 // despite the allow above.
2206 #[warn(missing_docs)]
2207 pub fn undocumented_two() -> int { 2 }
2210 // Missing documentation signals a warning here
2211 pub fn undocumented_too() -> int { 3 }
2215 This example shows how one can use `forbid` to disallow uses of `allow` for
2219 #[forbid(missing_docs)]
2221 // Attempting to toggle warning signals an error here
2222 #[allow(missing_docs)]
2224 pub fn undocumented_too() -> int { 2 }
2230 Some primitive Rust operations are defined in Rust code, rather than being
2231 implemented directly in C or assembly language. The definitions of these
2232 operations have to be easy for the compiler to find. The `lang` attribute
2233 makes it possible to declare these operations. For example, the `str` module
2234 in the Rust standard library defines the string equality function:
2238 pub fn eq_slice(a: &str, b: &str) -> bool {
2243 The name `str_eq` has a special meaning to the Rust compiler, and the presence
2244 of this definition means that it will use this definition when generating calls
2245 to the string equality function.
2247 A complete list of the built-in language items follows:
2249 #### Built-in Traits
2252 : Types that do not move ownership when used by-value.
2256 : Able to be sent across thread boundaries.
2258 : Has a size known at compile time.
2260 : Able to be safely shared between threads when aliased.
2264 These language items are traits:
2267 : Elements can be added (for example, integers and floats).
2269 : Elements can be subtracted.
2271 : Elements can be multiplied.
2273 : Elements have a division operation.
2275 : Elements have a remainder operation.
2277 : Elements can be negated arithmetically.
2279 : Elements can be negated logically.
2281 : Elements have an exclusive-or operation.
2283 : Elements have a bitwise `and` operation.
2285 : Elements have a bitwise `or` operation.
2287 : Elements have a left shift operation.
2289 : Elements have a right shift operation.
2291 : Elements can be indexed.
2293 : ___Needs filling in___
2295 : Elements can be compared for equality.
2297 : Elements have a partial ordering.
2299 : `*` can be applied, yielding a reference to another type
2301 : `*` can be applied, yielding a mutable reference to another type
2303 These are functions:
2306 : ___Needs filling in___
2308 : ___Needs filling in___
2310 : ___Needs filling in___
2312 : Compare two strings (`&str`) for equality.
2314 : Return a new unique string
2315 containing a copy of the contents of a unique string.
2320 : The type returned by the `type_id` intrinsic.
2322 : A type whose contents can be mutated through an immutable reference
2326 These types help drive the compiler's analysis
2329 : ___Needs filling in___
2331 : This type does not implement "copy", even if eligible
2333 : This type does not implement "send", even if eligible
2335 : This type does not implement "sync", even if eligible
2337 : ___Needs filling in___
2339 : Free memory that was allocated on the exchange heap.
2341 : Allocate memory on the exchange heap.
2342 * `closure_exchange_malloc`
2343 : ___Needs filling in___
2345 : Abort the program with an error.
2346 * `fail_bounds_check`
2347 : Abort the program with a bounds check error.
2349 : Free memory that was allocated on the managed heap.
2351 : ___Needs filling in___
2353 : ___Needs filling in___
2355 : ___Needs filling in___
2356 * `contravariant_lifetime`
2357 : The lifetime parameter should be considered contravariant
2358 * `covariant_lifetime`
2359 : The lifetime parameter should be considered covariant
2360 * `invariant_lifetime`
2361 : The lifetime parameter should be considered invariant
2363 : Allocate memory on the managed heap.
2365 : ___Needs filling in___
2367 : ___Needs filling in___
2369 : ___Needs filling in___
2370 * `contravariant_type`
2371 : The type parameter should be considered contravariant
2373 : The type parameter should be considered covariant
2375 : The type parameter should be considered invariant
2377 : ___Needs filling in___
2379 > **Note:** This list is likely to become out of date. We should auto-generate
2380 > it from `librustc/middle/lang_items.rs`.
2382 ### Inline attributes
2384 The inline attribute is used to suggest to the compiler to perform an inline
2385 expansion and place a copy of the function or static in the caller rather than
2386 generating code to call the function or access the static where it is defined.
2388 The compiler automatically inlines functions based on internal heuristics.
2389 Incorrectly inlining functions can actually making the program slower, so it
2390 should be used with care.
2392 Immutable statics are always considered inlineable unless marked with
2393 `#[inline(never)]`. It is undefined whether two different inlineable statics
2394 have the same memory address. In other words, the compiler is free to collapse
2395 duplicate inlineable statics together.
2397 `#[inline]` and `#[inline(always)]` always causes the function to be serialized
2398 into crate metadata to allow cross-crate inlining.
2400 There are three different types of inline attributes:
2402 * `#[inline]` hints the compiler to perform an inline expansion.
2403 * `#[inline(always)]` asks the compiler to always perform an inline expansion.
2404 * `#[inline(never)]` asks the compiler to never perform an inline expansion.
2408 The `deriving` attribute allows certain traits to be automatically implemented
2409 for data structures. For example, the following will create an `impl` for the
2410 `PartialEq` and `Clone` traits for `Foo`, the type parameter `T` will be given
2411 the `PartialEq` or `Clone` constraints for the appropriate `impl`:
2414 #[deriving(PartialEq, Clone)]
2421 The generated `impl` for `PartialEq` is equivalent to
2424 # struct Foo<T> { a: int, b: T }
2425 impl<T: PartialEq> PartialEq for Foo<T> {
2426 fn eq(&self, other: &Foo<T>) -> bool {
2427 self.a == other.a && self.b == other.b
2430 fn ne(&self, other: &Foo<T>) -> bool {
2431 self.a != other.a || self.b != other.b
2436 Supported traits for `deriving` are:
2438 * Comparison traits: `PartialEq`, `Eq`, `PartialOrd`, `Ord`.
2439 * Serialization: `Encodable`, `Decodable`. These require `serialize`.
2440 * `Clone`, to create `T` from `&T` via a copy.
2441 * `Default`, to create an empty instance of a data type.
2442 * `FromPrimitive`, to create an instance from a numeric primitive.
2443 * `Hash`, to iterate over the bytes in a data type.
2444 * `Rand`, to create a random instance of a data type.
2445 * `Show`, to format a value using the `{}` formatter.
2446 * `Zero`, to create a zero instance of a numeric data type.
2450 One can indicate the stability of an API using the following attributes:
2452 * `deprecated`: This item should no longer be used, e.g. it has been
2453 replaced. No guarantee of backwards-compatibility.
2454 * `experimental`: This item was only recently introduced or is
2455 otherwise in a state of flux. It may change significantly, or even
2456 be removed. No guarantee of backwards-compatibility.
2457 * `unstable`: This item is still under development, but requires more
2458 testing to be considered stable. No guarantee of backwards-compatibility.
2459 * `stable`: This item is considered stable, and will not change
2460 significantly. Guarantee of backwards-compatibility.
2461 * `frozen`: This item is very stable, and is unlikely to
2462 change. Guarantee of backwards-compatibility.
2463 * `locked`: This item will never change unless a serious bug is
2464 found. Guarantee of backwards-compatibility.
2466 These levels are directly inspired by
2467 [Node.js' "stability index"](http://nodejs.org/api/documentation.html).
2469 Stability levels are inherited, so an item's stability attribute is the default
2470 stability for everything nested underneath it.
2472 There are lints for disallowing items marked with certain levels: `deprecated`,
2473 `experimental` and `unstable`. For now, only `deprecated` warns by default, but
2474 this will change once the standard library has been stabilized. Stability
2475 levels are meant to be promises at the crate level, so these lints only apply
2476 when referencing items from an _external_ crate, not to items defined within
2477 the current crate. Items with no stability level are considered to be unstable
2478 for the purposes of the lint. One can give an optional string that will be
2479 displayed when the lint flags the use of an item.
2481 For example, if we define one crate called `stability_levels`:
2484 #[deprecated="replaced by `best`"]
2486 // delete everything
2490 // delete fewer things
2499 then the lints will work as follows for a client crate:
2503 extern crate stability_levels;
2504 use stability_levels::{bad, better, best};
2507 bad(); // "warning: use of deprecated item: replaced by `best`"
2509 better(); // "warning: use of unmarked item"
2511 best(); // no warning
2515 > **Note:** Currently these are only checked when applied to individual
2516 > functions, structs, methods and enum variants, *not* to entire modules,
2517 > traits, impls or enums themselves.
2519 ### Compiler Features
2521 Certain aspects of Rust may be implemented in the compiler, but they're not
2522 necessarily ready for every-day use. These features are often of "prototype
2523 quality" or "almost production ready", but may not be stable enough to be
2524 considered a full-fledged language feature.
2526 For this reason, Rust recognizes a special crate-level attribute of the form:
2529 #![feature(feature1, feature2, feature3)]
2532 This directive informs the compiler that the feature list: `feature1`,
2533 `feature2`, and `feature3` should all be enabled. This is only recognized at a
2534 crate-level, not at a module-level. Without this directive, all features are
2535 considered off, and using the features will result in a compiler error.
2537 The currently implemented features of the reference compiler are:
2539 * `asm` - The `asm!` macro provides a means for inline assembly. This is often
2540 useful, but the exact syntax for this feature along with its
2541 semantics are likely to change, so this macro usage must be opted
2544 * `concat_idents` - Allows use of the `concat_idents` macro, which is in many
2545 ways insufficient for concatenating identifiers, and may be
2546 removed entirely for something more wholesome.
2548 * `default_type_params` - Allows use of default type parameters. The future of
2549 this feature is uncertain.
2551 * `intrinsics` - Allows use of the "rust-intrinsics" ABI. Compiler intrinsics
2552 are inherently unstable and no promise about them is made.
2554 * `lang_items` - Allows use of the `#[lang]` attribute. Like `intrinsics`,
2555 lang items are inherently unstable and no promise about them
2558 * `link_args` - This attribute is used to specify custom flags to the linker,
2559 but usage is strongly discouraged. The compiler's usage of the
2560 system linker is not guaranteed to continue in the future, and
2561 if the system linker is not used then specifying custom flags
2562 doesn't have much meaning.
2564 * `link_llvm_intrinsics` – Allows linking to LLVM intrinsics via
2565 `#[link_name="llvm.*"]`.
2567 * `linkage` - Allows use of the `linkage` attribute, which is not portable.
2569 * `log_syntax` - Allows use of the `log_syntax` macro attribute, which is a
2570 nasty hack that will certainly be removed.
2572 * `macro_rules` - The definition of new macros. This does not encompass
2573 macro-invocation, that is always enabled by default, this
2574 only covers the definition of new macros. There are currently
2575 various problems with invoking macros, how they interact with
2576 their environment, and possibly how they are used outside of
2577 location in which they are defined. Macro definitions are
2578 likely to change slightly in the future, so they are
2579 currently hidden behind this feature.
2581 * `non_ascii_idents` - The compiler supports the use of non-ascii identifiers,
2582 but the implementation is a little rough around the
2583 edges, so this can be seen as an experimental feature
2584 for now until the specification of identifiers is fully
2587 * `once_fns` - Onceness guarantees a closure is only executed once. Defining a
2588 closure as `once` is unlikely to be supported going forward. So
2589 they are hidden behind this feature until they are to be removed.
2591 * `phase` - Usage of the `#[phase]` attribute allows loading compiler plugins
2592 for custom lints or syntax extensions. The implementation is
2593 considered unwholesome and in need of overhaul, and it is not clear
2594 what they will look like moving forward.
2596 * `plugin_registrar` - Indicates that a crate has [compiler plugins][plugin] that it
2597 wants to load. As with `phase`, the implementation is
2598 in need of an overhaul, and it is not clear that plugins
2599 defined using this will continue to work.
2601 * `quote` - Allows use of the `quote_*!` family of macros, which are
2602 implemented very poorly and will likely change significantly
2603 with a proper implementation.
2605 * `rustc_diagnostic_macros`- A mysterious feature, used in the implementation
2606 of rustc, not meant for mortals.
2608 * `simd` - Allows use of the `#[simd]` attribute, which is overly simple and
2609 not the SIMD interface we want to expose in the long term.
2611 * `struct_inherit` - Allows using struct inheritance, which is barely
2612 implemented and will probably be removed. Don't use this.
2614 * `struct_variant` - Structural enum variants (those with named fields). It is
2615 currently unknown whether this style of enum variant is as
2616 fully supported as the tuple-forms, and it's not certain
2617 that this style of variant should remain in the language.
2618 For now this style of variant is hidden behind a feature
2621 * `thread_local` - The usage of the `#[thread_local]` attribute is experimental
2622 and should be seen as unstable. This attribute is used to
2623 declare a `static` as being unique per-thread leveraging
2624 LLVM's implementation which works in concert with the kernel
2625 loader and dynamic linker. This is not necessarily available
2626 on all platforms, and usage of it is discouraged (rust
2627 focuses more on thread-local data instead of thread-local
2630 * `trace_macros` - Allows use of the `trace_macros` macro, which is a nasty
2631 hack that will certainly be removed.
2633 * `unboxed_closures` - Rust's new closure design, which is currently a work in
2634 progress feature with many known bugs.
2636 * `unsafe_destructor` - Allows use of the `#[unsafe_destructor]` attribute,
2637 which is considered wildly unsafe and will be
2638 obsoleted by language improvements.
2640 * `associated_types` - Allows type aliases in traits. Experimental.
2642 If a feature is promoted to a language feature, then all existing programs will
2643 start to receive compilation warnings about #[feature] directives which enabled
2644 the new feature (because the directive is no longer necessary). However, if a
2645 feature is decided to be removed from the language, errors will be issued (if
2646 there isn't a parser error first). The directive in this case is no longer
2647 necessary, and it's likely that existing code will break if the feature isn't
2650 If an unknown feature is found in a directive, it results in a compiler error.
2651 An unknown feature is one which has never been recognized by the compiler.
2653 # Statements and expressions
2655 Rust is _primarily_ an expression language. This means that most forms of
2656 value-producing or effect-causing evaluation are directed by the uniform syntax
2657 category of _expressions_. Each kind of expression can typically _nest_ within
2658 each other kind of expression, and rules for evaluation of expressions involve
2659 specifying both the value produced by the expression and the order in which its
2660 sub-expressions are themselves evaluated.
2662 In contrast, statements in Rust serve _mostly_ to contain and explicitly
2663 sequence expression evaluation.
2667 A _statement_ is a component of a block, which is in turn a component of an
2668 outer [expression](#expressions) or [function](#functions).
2670 Rust has two kinds of statement: [declaration
2671 statements](#declaration-statements) and [expression
2672 statements](#expression-statements).
2674 ### Declaration statements
2676 A _declaration statement_ is one that introduces one or more *names* into the
2677 enclosing statement block. The declared names may denote new slots or new
2680 #### Item declarations
2682 An _item declaration statement_ has a syntactic form identical to an
2683 [item](#items) declaration within a module. Declaring an item — a
2684 function, enumeration, structure, type, static, trait, implementation or module
2685 — locally within a statement block is simply a way of restricting its
2686 scope to a narrow region containing all of its uses; it is otherwise identical
2687 in meaning to declaring the item outside the statement block.
2689 > **Note**: there is no implicit capture of the function's dynamic environment when
2690 > declaring a function-local item.
2692 #### Slot declarations
2695 let_decl : "let" pat [':' type ] ? [ init ] ? ';' ;
2696 init : [ '=' ] expr ;
2699 A _slot declaration_ introduces a new set of slots, given by a pattern. The
2700 pattern may be followed by a type annotation, and/or an initializer expression.
2701 When no type annotation is given, the compiler will infer the type, or signal
2702 an error if insufficient type information is available for definite inference.
2703 Any slots introduced by a slot declaration are visible from the point of
2704 declaration until the end of the enclosing block scope.
2706 ### Expression statements
2708 An _expression statement_ is one that evaluates an [expression](#expressions)
2709 and ignores its result. The type of an expression statement `e;` is always
2710 `()`, regardless of the type of `e`. As a rule, an expression statement's
2711 purpose is to trigger the effects of evaluating its expression.
2715 An expression may have two roles: it always produces a *value*, and it may have
2716 *effects* (otherwise known as "side effects"). An expression *evaluates to* a
2717 value, and has effects during *evaluation*. Many expressions contain
2718 sub-expressions (operands). The meaning of each kind of expression dictates
2721 * Whether or not to evaluate the sub-expressions when evaluating the expression
2722 * The order in which to evaluate the sub-expressions
2723 * How to combine the sub-expressions' values to obtain the value of the expression
2725 In this way, the structure of expressions dictates the structure of execution.
2726 Blocks are just another kind of expression, so blocks, statements, expressions,
2727 and blocks again can recursively nest inside each other to an arbitrary depth.
2729 #### Lvalues, rvalues and temporaries
2731 Expressions are divided into two main categories: _lvalues_ and _rvalues_.
2732 Likewise within each expression, sub-expressions may occur in _lvalue context_
2733 or _rvalue context_. The evaluation of an expression depends both on its own
2734 category and the context it occurs within.
2736 An lvalue is an expression that represents a memory location. These expressions
2737 are [paths](#path-expressions) (which refer to local variables, function and
2738 method arguments, or static variables), dereferences (`*expr`), [indexing
2739 expressions](#index-expressions) (`expr[expr]`), and [field
2740 references](#field-expressions) (`expr.f`). All other expressions are rvalues.
2742 The left operand of an [assignment](#assignment-expressions) or
2743 [compound-assignment](#compound-assignment-expressions) expression is an lvalue
2744 context, as is the single operand of a unary
2745 [borrow](#unary-operator-expressions). All other expression contexts are
2748 When an lvalue is evaluated in an _lvalue context_, it denotes a memory
2749 location; when evaluated in an _rvalue context_, it denotes the value held _in_
2750 that memory location.
2752 When an rvalue is used in an lvalue context, a temporary un-named lvalue is
2753 created and used instead. A temporary's lifetime equals the largest lifetime
2754 of any reference that points to it.
2756 #### Moved and copied types
2758 When a [local variable](#memory-slots) is used as an
2759 [rvalue](#lvalues,-rvalues-and-temporaries) the variable will either be moved
2760 or copied, depending on its type. For types that contain [owning
2761 pointers](#pointer-types) or values that implement the special trait `Drop`,
2762 the variable is moved. All other types are copied.
2764 ### Literal expressions
2766 A _literal expression_ consists of one of the [literal](#literals) forms
2767 described earlier. It directly describes a number, character, string, boolean
2768 value, or the unit value.
2772 "hello"; // string type
2773 '5'; // character type
2777 ### Path expressions
2779 A [path](#paths) used as an expression context denotes either a local variable
2780 or an item. Path expressions are [lvalues](#lvalues,-rvalues-and-temporaries).
2782 ### Tuple expressions
2784 Tuples are written by enclosing zero or more comma-separated expressions in
2785 parentheses. They are used to create [tuple-typed](#tuple-types) values.
2793 ### Unit expressions
2795 The expression `()` denotes the _unit value_, the only value of the type with
2798 ### Structure expressions
2801 struct_expr : expr_path '{' ident ':' expr
2802 [ ',' ident ':' expr ] *
2805 [ ',' expr ] * ')' |
2809 There are several forms of structure expressions. A _structure expression_
2810 consists of the [path](#paths) of a [structure item](#structures), followed by
2811 a brace-enclosed list of one or more comma-separated name-value pairs,
2812 providing the field values of a new instance of the structure. A field name
2813 can be any identifier, and is separated from its value expression by a colon.
2814 The location denoted by a structure field is mutable if and only if the
2815 enclosing structure is mutable.
2817 A _tuple structure expression_ consists of the [path](#paths) of a [structure
2818 item](#structures), followed by a parenthesized list of one or more
2819 comma-separated expressions (in other words, the path of a structure item
2820 followed by a tuple expression). The structure item must be a tuple structure
2823 A _unit-like structure expression_ consists only of the [path](#paths) of a
2824 [structure item](#structures).
2826 The following are examples of structure expressions:
2829 # struct Point { x: f64, y: f64 }
2830 # struct TuplePoint(f64, f64);
2831 # mod game { pub struct User<'a> { pub name: &'a str, pub age: uint, pub score: uint } }
2832 # struct Cookie; fn some_fn<T>(t: T) {}
2833 Point {x: 10.0, y: 20.0};
2834 TuplePoint(10.0, 20.0);
2835 let u = game::User {name: "Joe", age: 35, score: 100_000};
2836 some_fn::<Cookie>(Cookie);
2839 A structure expression forms a new value of the named structure type. Note
2840 that for a given *unit-like* structure type, this will always be the same
2843 A structure expression can terminate with the syntax `..` followed by an
2844 expression to denote a functional update. The expression following `..` (the
2845 base) must have the same structure type as the new structure type being formed.
2846 The entire expression denotes the result of constructing a new structure (with
2847 the same type as the base expression) with the given values for the fields that
2848 were explicitly specified and the values in the base expression for all other
2852 # struct Point3d { x: int, y: int, z: int }
2853 let base = Point3d {x: 1, y: 2, z: 3};
2854 Point3d {y: 0, z: 10, .. base};
2857 ### Block expressions
2860 block_expr : '{' [ view_item ] *
2861 [ stmt ';' | item ] *
2865 A _block expression_ is similar to a module in terms of the declarations that
2866 are possible. Each block conceptually introduces a new namespace scope. View
2867 items can bring new names into scopes and declared items are in scope for only
2870 A block will execute each statement sequentially, and then execute the
2871 expression (if given). If the final expression is omitted, the type and return
2872 value of the block are `()`, but if it is provided, the type and return value
2873 of the block are that of the expression itself.
2875 ### Method-call expressions
2878 method_call_expr : expr '.' ident paren_expr_list ;
2881 A _method call_ consists of an expression followed by a single dot, an
2882 identifier, and a parenthesized expression-list. Method calls are resolved to
2883 methods on specific traits, either statically dispatching to a method if the
2884 exact `self`-type of the left-hand-side is known, or dynamically dispatching if
2885 the left-hand-side expression is an indirect [object type](#object-types).
2887 ### Field expressions
2890 field_expr : expr '.' ident ;
2893 A _field expression_ consists of an expression followed by a single dot and an
2894 identifier, when not immediately followed by a parenthesized expression-list
2895 (the latter is a [method call expression](#method-call-expressions)). A field
2896 expression denotes a field of a [structure](#structure-types).
2901 (Struct {a: 10, b: 20}).a;
2904 A field access is an [lvalue](#lvalues,-rvalues-and-temporaries) referring to
2905 the value of that field. When the type providing the field inherits mutability,
2906 it can be [assigned](#assignment-expressions) to.
2908 Also, if the type of the expression to the left of the dot is a pointer, it is
2909 automatically dereferenced to make the field access possible.
2911 ### Array expressions
2914 array_expr : '[' "mut" ? vec_elems? ']' ;
2916 array_elems : [expr [',' expr]*] | [expr ',' ".." expr] ;
2919 An [array](#array,-and-slice-types) _expression_ is written by enclosing zero
2920 or more comma-separated expressions of uniform type in square brackets.
2922 In the `[expr ',' ".." expr]` form, the expression after the `".."` must be a
2923 constant expression that can be evaluated at compile time, such as a
2924 [literal](#literals) or a [static item](#static-items).
2928 ["a", "b", "c", "d"];
2929 [0i; 128]; // array with 128 zeros
2930 [0u8, 0u8, 0u8, 0u8];
2933 ### Index expressions
2936 idx_expr : expr '[' expr ']' ;
2939 [Array](#array,-and-slice-types)-typed expressions can be indexed by
2940 writing a square-bracket-enclosed expression (the index) after them. When the
2941 array is mutable, the resulting [lvalue](#lvalues,-rvalues-and-temporaries) can
2944 Indices are zero-based, and may be of any integral type. Vector access is
2945 bounds-checked at run-time. When the check fails, it will put the thread in a
2950 (["a", "b"])[10]; // panics
2953 ### Unary operator expressions
2955 Rust defines six symbolic unary operators. They are all written as prefix
2956 operators, before the expression they apply to.
2959 : Negation. May only be applied to numeric types.
2961 : Dereference. When applied to a [pointer](#pointer-types) it denotes the
2962 pointed-to location. For pointers to mutable locations, the resulting
2963 [lvalue](#lvalues,-rvalues-and-temporaries) can be assigned to.
2964 On non-pointer types, it calls the `deref` method of the `std::ops::Deref`
2965 trait, or the `deref_mut` method of the `std::ops::DerefMut` trait (if
2966 implemented by the type and required for an outer expression that will or
2967 could mutate the dereference), and produces the result of dereferencing the
2968 `&` or `&mut` borrowed pointer returned from the overload method.
2971 : Logical negation. On the boolean type, this flips between `true` and
2972 `false`. On integer types, this inverts the individual bits in the
2973 two's complement representation of the value.
2975 : [Boxing](#pointer-types) operators. Allocate a box to hold the value they
2976 are applied to, and store the value in it. `box` creates a box.
2978 : Borrow operator. Returns a reference, pointing to its operand. The operand
2979 of a borrow is statically proven to outlive the resulting pointer. If the
2980 borrow-checker cannot prove this, it is a compilation error.
2982 ### Binary operator expressions
2985 binop_expr : expr binop expr ;
2988 Binary operators expressions are given in terms of [operator
2989 precedence](#operator-precedence).
2991 #### Arithmetic operators
2993 Binary arithmetic expressions are syntactic sugar for calls to built-in traits,
2994 defined in the `std::ops` module of the `std` library. This means that
2995 arithmetic operators can be overridden for user-defined types. The default
2996 meaning of the operators on standard types is given here.
2999 : Addition and array/string concatenation.
3000 Calls the `add` method on the `std::ops::Add` trait.
3003 Calls the `sub` method on the `std::ops::Sub` trait.
3006 Calls the `mul` method on the `std::ops::Mul` trait.
3009 Calls the `div` method on the `std::ops::Div` trait.
3012 Calls the `rem` method on the `std::ops::Rem` trait.
3014 #### Bitwise operators
3016 Like the [arithmetic operators](#arithmetic-operators), bitwise operators are
3017 syntactic sugar for calls to methods of built-in traits. This means that
3018 bitwise operators can be overridden for user-defined types. The default
3019 meaning of the operators on standard types is given here.
3023 Calls the `bitand` method of the `std::ops::BitAnd` trait.
3026 Calls the `bitor` method of the `std::ops::BitOr` trait.
3029 Calls the `bitxor` method of the `std::ops::BitXor` trait.
3031 : Logical left shift.
3032 Calls the `shl` method of the `std::ops::Shl` trait.
3034 : Logical right shift.
3035 Calls the `shr` method of the `std::ops::Shr` trait.
3037 #### Lazy boolean operators
3039 The operators `||` and `&&` may be applied to operands of boolean type. The
3040 `||` operator denotes logical 'or', and the `&&` operator denotes logical
3041 'and'. They differ from `|` and `&` in that the right-hand operand is only
3042 evaluated when the left-hand operand does not already determine the result of
3043 the expression. That is, `||` only evaluates its right-hand operand when the
3044 left-hand operand evaluates to `false`, and `&&` only when it evaluates to
3047 #### Comparison operators
3049 Comparison operators are, like the [arithmetic
3050 operators](#arithmetic-operators), and [bitwise operators](#bitwise-operators),
3051 syntactic sugar for calls to built-in traits. This means that comparison
3052 operators can be overridden for user-defined types. The default meaning of the
3053 operators on standard types is given here.
3057 Calls the `eq` method on the `std::cmp::PartialEq` trait.
3060 Calls the `ne` method on the `std::cmp::PartialEq` trait.
3063 Calls the `lt` method on the `std::cmp::PartialOrd` trait.
3066 Calls the `gt` method on the `std::cmp::PartialOrd` trait.
3068 : Less than or equal.
3069 Calls the `le` method on the `std::cmp::PartialOrd` trait.
3071 : Greater than or equal.
3072 Calls the `ge` method on the `std::cmp::PartialOrd` trait.
3074 #### Type cast expressions
3076 A type cast expression is denoted with the binary operator `as`.
3078 Executing an `as` expression casts the value on the left-hand side to the type
3079 on the right-hand side.
3081 A numeric value can be cast to any numeric type. A raw pointer value can be
3082 cast to or from any integral type or raw pointer type. Any other cast is
3083 unsupported and will fail to compile.
3085 An example of an `as` expression:
3088 # fn sum(v: &[f64]) -> f64 { 0.0 }
3089 # fn len(v: &[f64]) -> int { 0 }
3091 fn avg(v: &[f64]) -> f64 {
3092 let sum: f64 = sum(v);
3093 let sz: f64 = len(v) as f64;
3098 #### Assignment expressions
3100 An _assignment expression_ consists of an
3101 [lvalue](#lvalues,-rvalues-and-temporaries) expression followed by an equals
3102 sign (`=`) and an [rvalue](#lvalues,-rvalues-and-temporaries) expression.
3104 Evaluating an assignment expression [either copies or
3105 moves](#moved-and-copied-types) its right-hand operand to its left-hand
3115 #### Compound assignment expressions
3117 The `+`, `-`, `*`, `/`, `%`, `&`, `|`, `^`, `<<`, and `>>` operators may be
3118 composed with the `=` operator. The expression `lval OP= val` is equivalent to
3119 `lval = lval OP val`. For example, `x = x + 1` may be written as `x += 1`.
3121 Any such expression always has the [`unit`](#primitive-types) type.
3123 #### Operator precedence
3125 The precedence of Rust binary operators is ordered as follows, going from
3128 ```{.text .precedence}
3143 Operators at the same precedence level are evaluated left-to-right. [Unary
3144 operators](#unary-operator-expressions) have the same precedence level and are
3145 stronger than any of the binary operators.
3147 ### Grouped expressions
3149 An expression enclosed in parentheses evaluates to the result of the enclosed
3150 expression. Parentheses can be used to explicitly specify evaluation order
3151 within an expression.
3154 paren_expr : '(' expr ')' ;
3157 An example of a parenthesized expression:
3160 let x: int = (2 + 3) * 4;
3164 ### Call expressions
3167 expr_list : [ expr [ ',' expr ]* ] ? ;
3168 paren_expr_list : '(' expr_list ')' ;
3169 call_expr : expr paren_expr_list ;
3172 A _call expression_ invokes a function, providing zero or more input slots and
3173 an optional reference slot to serve as the function's output, bound to the
3174 `lval` on the right hand side of the call. If the function eventually returns,
3175 then the expression completes.
3177 Some examples of call expressions:
3180 # fn add(x: int, y: int) -> int { 0 }
3182 let x: int = add(1, 2);
3183 let pi: Option<f32> = "3.14".parse();
3186 ### Lambda expressions
3189 ident_list : [ ident [ ',' ident ]* ] ? ;
3190 lambda_expr : '|' ident_list '|' expr ;
3193 A _lambda expression_ (sometimes called an "anonymous function expression")
3194 defines a function and denotes it as a value, in a single expression. A lambda
3195 expression is a pipe-symbol-delimited (`|`) list of identifiers followed by an
3198 A lambda expression denotes a function that maps a list of parameters
3199 (`ident_list`) onto the expression that follows the `ident_list`. The
3200 identifiers in the `ident_list` are the parameters to the function. These
3201 parameters' types need not be specified, as the compiler infers them from
3204 Lambda expressions are most useful when passing functions as arguments to other
3205 functions, as an abbreviation for defining and capturing a separate function.
3207 Significantly, lambda expressions _capture their environment_, which regular
3208 [function definitions](#functions) do not. The exact type of capture depends
3209 on the [function type](#function-types) inferred for the lambda expression. In
3210 the simplest and least-expensive form (analogous to a ```|| { }``` expression),
3211 the lambda expression captures its environment by reference, effectively
3212 borrowing pointers to all outer variables mentioned inside the function.
3213 Alternately, the compiler may infer that a lambda expression should copy or
3214 move values (depending on their type.) from the environment into the lambda
3215 expression's captured environment.
3217 In this example, we define a function `ten_times` that takes a higher-order
3218 function argument, and call it with a lambda expression as an argument.
3221 fn ten_times(f: |int|) {
3229 ten_times(|j| println!("hello, {}", j));
3235 while_expr : "while" no_struct_literal_expr '{' block '}' ;
3238 A `while` loop begins by evaluating the boolean loop conditional expression.
3239 If the loop conditional expression evaluates to `true`, the loop body block
3240 executes and control returns to the loop conditional expression. If the loop
3241 conditional expression evaluates to `false`, the `while` expression completes.
3256 A `loop` expression denotes an infinite loop.
3259 loop_expr : [ lifetime ':' ] "loop" '{' block '}';
3262 A `loop` expression may optionally have a _label_. If a label is present, then
3263 labeled `break` and `continue` expressions nested within this loop may exit out
3264 of this loop or return control to its head. See [Break
3265 expressions](#break-expressions) and [Continue
3266 expressions](#continue-expressions).
3268 ### Break expressions
3271 break_expr : "break" [ lifetime ];
3274 A `break` expression has an optional _label_. If the label is absent, then
3275 executing a `break` expression immediately terminates the innermost loop
3276 enclosing it. It is only permitted in the body of a loop. If the label is
3277 present, then `break foo` terminates the loop with label `foo`, which need not
3278 be the innermost label enclosing the `break` expression, but must enclose it.
3280 ### Continue expressions
3283 continue_expr : "continue" [ lifetime ];
3286 A `continue` expression has an optional _label_. If the label is absent, then
3287 executing a `continue` expression immediately terminates the current iteration
3288 of the innermost loop enclosing it, returning control to the loop *head*. In
3289 the case of a `while` loop, the head is the conditional expression controlling
3290 the loop. In the case of a `for` loop, the head is the call-expression
3291 controlling the loop. If the label is present, then `continue foo` returns
3292 control to the head of the loop with label `foo`, which need not be the
3293 innermost label enclosing the `break` expression, but must enclose it.
3295 A `continue` expression is only permitted in the body of a loop.
3300 for_expr : "for" pat "in" no_struct_literal_expr '{' block '}' ;
3303 A `for` expression is a syntactic construct for looping over elements provided
3304 by an implementation of `std::iter::Iterator`.
3306 An example of a for loop over the contents of an array:
3310 # fn bar(f: Foo) { }
3315 let v: &[Foo] = &[a, b, c];
3322 An example of a for loop over a series of integers:
3325 # fn bar(b:uint) { }
3326 for i in range(0u, 256) {
3334 if_expr : "if" no_struct_literal_expr '{' block '}'
3337 else_tail : "else" [ if_expr | if_let_expr
3341 An `if` expression is a conditional branch in program control. The form of an
3342 `if` expression is a condition expression, followed by a consequent block, any
3343 number of `else if` conditions and blocks, and an optional trailing `else`
3344 block. The condition expressions must have type `bool`. If a condition
3345 expression evaluates to `true`, the consequent block is executed and any
3346 subsequent `else if` or `else` block is skipped. If a condition expression
3347 evaluates to `false`, the consequent block is skipped and any subsequent `else
3348 if` condition is evaluated. If all `if` and `else if` conditions evaluate to
3349 `false` then any `else` block is executed.
3351 ### Match expressions
3354 match_expr : "match" no_struct_literal_expr '{' match_arm * '}' ;
3356 match_arm : attribute * match_pat "=>" [ expr "," | '{' block '}' ] ;
3358 match_pat : pat [ '|' pat ] * [ "if" expr ] ? ;
3361 A `match` expression branches on a *pattern*. The exact form of matching that
3362 occurs depends on the pattern. Patterns consist of some combination of
3363 literals, destructured arrays or enum constructors, structures and tuples,
3364 variable binding specifications, wildcards (`..`), and placeholders (`_`). A
3365 `match` expression has a *head expression*, which is the value to compare to
3366 the patterns. The type of the patterns must equal the type of the head
3369 In a pattern whose head expression has an `enum` type, a placeholder (`_`)
3370 stands for a *single* data field, whereas a wildcard `..` stands for *all* the
3371 fields of a particular variant. For example:
3374 enum List<X> { Nil, Cons(X, Box<List<X>>) }
3376 let x: List<int> = List::Cons(10, box List::Cons(11, box List::Nil));
3379 List::Cons(_, box List::Nil) => panic!("singleton list"),
3380 List::Cons(..) => return,
3381 List::Nil => panic!("empty list")
3385 The first pattern matches lists constructed by applying `Cons` to any head
3386 value, and a tail value of `box Nil`. The second pattern matches _any_ list
3387 constructed with `Cons`, ignoring the values of its arguments. The difference
3388 between `_` and `..` is that the pattern `C(_)` is only type-correct if `C` has
3389 exactly one argument, while the pattern `C(..)` is type-correct for any enum
3390 variant `C`, regardless of how many arguments `C` has.
3392 Used inside an array pattern, `..` stands for any number of elements, when the
3393 `advanced_slice_patterns` feature gate is turned on. This wildcard can be used
3394 at most once for a given array, which implies that it cannot be used to
3395 specifically match elements that are at an unknown distance from both ends of a
3396 array, like `[.., 42, ..]`. If followed by a variable name, it will bind the
3397 corresponding slice to the variable. Example:
3400 # #![feature(advanced_slice_patterns)]
3401 fn is_symmetric(list: &[uint]) -> bool {
3404 [x, inside.., y] if x == y => is_symmetric(inside),
3410 let sym = &[0, 1, 4, 2, 4, 1, 0];
3411 let not_sym = &[0, 1, 7, 2, 4, 1, 0];
3412 assert!(is_symmetric(sym));
3413 assert!(!is_symmetric(not_sym));
3417 A `match` behaves differently depending on whether or not the head expression
3418 is an [lvalue or an rvalue](#lvalues,-rvalues-and-temporaries). If the head
3419 expression is an rvalue, it is first evaluated into a temporary location, and
3420 the resulting value is sequentially compared to the patterns in the arms until
3421 a match is found. The first arm with a matching pattern is chosen as the branch
3422 target of the `match`, any variables bound by the pattern are assigned to local
3423 variables in the arm's block, and control enters the block.
3425 When the head expression is an lvalue, the match does not allocate a temporary
3426 location (however, a by-value binding may copy or move from the lvalue). When
3427 possible, it is preferable to match on lvalues, as the lifetime of these
3428 matches inherits the lifetime of the lvalue, rather than being restricted to
3429 the inside of the match.
3431 An example of a `match` expression:
3434 # fn process_pair(a: int, b: int) { }
3435 # fn process_ten() { }
3437 enum List<X> { Nil, Cons(X, Box<List<X>>) }
3439 let x: List<int> = List::Cons(10, box List::Cons(11, box List::Nil));
3442 List::Cons(a, box List::Cons(b, _)) => {
3445 List::Cons(10, _) => {
3457 Patterns that bind variables default to binding to a copy or move of the
3458 matched value (depending on the matched value's type). This can be changed to
3459 bind to a reference by using the `ref` keyword, or to a mutable reference using
3462 Subpatterns can also be bound to variables by the use of the syntax `variable @
3463 subpattern`. For example:
3466 enum List { Nil, Cons(uint, Box<List>) }
3468 fn is_sorted(list: &List) -> bool {
3470 List::Nil | List::Cons(_, box List::Nil) => true,
3471 List::Cons(x, ref r @ box List::Cons(_, _)) => {
3473 box List::Cons(y, _) => (x <= y) && is_sorted(&**r),
3481 let a = List::Cons(6, box List::Cons(7, box List::Cons(42, box List::Nil)));
3482 assert!(is_sorted(&a));
3487 Patterns can also dereference pointers by using the `&`, `box` symbols,
3488 as appropriate. For example, these two matches on `x: &int` are equivalent:
3492 let y = match *x { 0 => "zero", _ => "some" };
3493 let z = match x { &0 => "zero", _ => "some" };
3498 A pattern that's just an identifier, like `Nil` in the previous example, could
3499 either refer to an enum variant that's in scope, or bind a new variable. The
3500 compiler resolves this ambiguity by forbidding variable bindings that occur in
3501 `match` patterns from shadowing names of variants that are in scope. For
3502 example, wherever `List` is in scope, a `match` pattern would not be able to
3503 bind `Nil` as a new name. The compiler interprets a variable pattern `x` as a
3504 binding _only_ if there is no variant named `x` in scope. A convention you can
3505 use to avoid conflicts is simply to name variants with upper-case letters, and
3506 local variables with lower-case letters.
3508 Multiple match patterns may be joined with the `|` operator. A range of values
3509 may be specified with `...`. For example:
3514 let message = match x {
3515 0 | 1 => "not many",
3521 Range patterns only work on scalar types (like integers and characters; not
3522 like arrays and structs, which have sub-components). A range pattern may not
3523 be a sub-range of another range pattern inside the same `match`.
3525 Finally, match patterns can accept *pattern guards* to further refine the
3526 criteria for matching a case. Pattern guards appear after the pattern and
3527 consist of a bool-typed expression following the `if` keyword. A pattern guard
3528 may refer to the variables bound within the pattern they follow.
3531 # let maybe_digit = Some(0);
3532 # fn process_digit(i: int) { }
3533 # fn process_other(i: int) { }
3535 let message = match maybe_digit {
3536 Some(x) if x < 10 => process_digit(x),
3537 Some(x) => process_other(x),
3542 ### If let expressions
3545 if_let_expr : "if" "let" pat '=' expr '{' block '}'
3547 else_tail : "else" [ if_expr | if_let_expr | '{' block '}' ] ;
3550 An `if let` expression is semantically identical to an `if` expression but in place
3551 of a condition expression it expects a refutable let statement. If the value of the
3552 expression on the right hand side of the let statement matches the pattern, the corresponding
3553 block will execute, otherwise flow proceeds to the first `else` block that follows.
3558 while_let_expr : "while" "let" pat '=' expr '{' block '}' ;
3561 A `while let` loop is semantically identical to a `while` loop but in place of a
3562 condition expression it expects a refutable let statement. If the value of the
3563 expression on the right hand side of the let statement matches the pattern, the
3564 loop body block executes and control returns to the pattern matching statement.
3565 Otherwise, the while expression completes.
3567 ### Return expressions
3570 return_expr : "return" expr ? ;
3573 Return expressions are denoted with the keyword `return`. Evaluating a `return`
3574 expression moves its argument into the output slot of the current function,
3575 destroys the current function activation frame, and transfers control to the
3578 An example of a `return` expression:
3581 fn max(a: int, b: int) -> int {
3593 Every slot, item and value in a Rust program has a type. The _type_ of a
3594 *value* defines the interpretation of the memory holding it.
3596 Built-in types and type-constructors are tightly integrated into the language,
3597 in nontrivial ways that are not possible to emulate in user-defined types.
3598 User-defined types have limited capabilities.
3602 The primitive types are the following:
3604 * The "unit" type `()`, having the single "unit" value `()` (occasionally called
3606 * The boolean type `bool` with values `true` and `false`.
3607 * The machine types.
3608 * The machine-dependent integer and floating-point types.
3610 [^unittype]: The "unit" value `()` is *not* a sentinel "null pointer" value for
3611 reference slots; the "unit" type is the implicit return type from functions
3612 otherwise lacking a return type, and can be used in other contexts (such as
3613 message-sending or type-parametric code) as a zero-size type.]
3617 The machine types are the following:
3619 * The unsigned word types `u8`, `u16`, `u32` and `u64`, with values drawn from
3620 the integer intervals [0, 2^8 - 1], [0, 2^16 - 1], [0, 2^32 - 1] and
3621 [0, 2^64 - 1] respectively.
3623 * The signed two's complement word types `i8`, `i16`, `i32` and `i64`, with
3624 values drawn from the integer intervals [-(2^(7)), 2^7 - 1],
3625 [-(2^(15)), 2^15 - 1], [-(2^(31)), 2^31 - 1], [-(2^(63)), 2^63 - 1]
3628 * The IEEE 754-2008 `binary32` and `binary64` floating-point types: `f32` and
3629 `f64`, respectively.
3631 #### Machine-dependent integer types
3633 The `uint` type is an unsigned integer type with the same number of bits as the
3634 platform's pointer type. It can represent every memory address in the process.
3636 The `int` type is a signed integer type with the same number of bits as the
3637 platform's pointer type. The theoretical upper bound on object and array size
3638 is the maximum `int` value. This ensures that `int` can be used to calculate
3639 differences between pointers into an object or array and can address every byte
3640 within an object along with one byte past the end.
3644 The types `char` and `str` hold textual data.
3646 A value of type `char` is a [Unicode scalar value](
3647 http://www.unicode.org/glossary/#unicode_scalar_value) (ie. a code point that
3648 is not a surrogate), represented as a 32-bit unsigned word in the 0x0000 to
3649 0xD7FF or 0xE000 to 0x10FFFF range. A `[char]` array is effectively an UCS-4 /
3652 A value of type `str` is a Unicode string, represented as an array of 8-bit
3653 unsigned bytes holding a sequence of UTF-8 codepoints. Since `str` is of
3654 unknown size, it is not a _first class_ type, but can only be instantiated
3655 through a pointer type, such as `&str` or `String`.
3659 A tuple *type* is a heterogeneous product of other types, called the *elements*
3660 of the tuple. It has no nominal name and is instead structurally typed.
3662 Tuple types and values are denoted by listing the types or values of their
3663 elements, respectively, in a parenthesized, comma-separated list.
3665 Because tuple elements don't have a name, they can only be accessed by
3668 The members of a tuple are laid out in memory contiguously, in order specified
3671 An example of a tuple type and its use:
3674 type Pair<'a> = (int, &'a str);
3675 let p: Pair<'static> = (10, "hello");
3677 assert!(b != "world");
3680 ### Array, and Slice types
3682 Rust has two different types for a list of items:
3684 * `[T ..N]`, an 'array'
3685 * `&[T]`, a 'slice'.
3687 An array has a fixed size, and can be allocated on either the stack or the
3690 A slice is a 'view' into an array. It doesn't own the data it points
3693 An example of each kind:
3696 let vec: Vec<int> = vec![1, 2, 3];
3697 let arr: [int; 3] = [1, 2, 3];
3698 let s: &[int] = vec.as_slice();
3701 As you can see, the `vec!` macro allows you to create a `Vec<T>` easily. The
3702 `vec!` macro is also part of the standard library, rather than the language.
3704 All in-bounds elements of arrays, and slices are always initialized, and access
3705 to an array or slice is always bounds-checked.
3709 A `struct` *type* is a heterogeneous product of other types, called the
3710 *fields* of the type.[^structtype]
3712 [^structtype]: `struct` types are analogous `struct` types in C,
3713 the *record* types of the ML family,
3714 or the *structure* types of the Lisp family.
3716 New instances of a `struct` can be constructed with a [struct
3717 expression](#structure-expressions).
3719 The memory layout of a `struct` is undefined by default to allow for compiler
3720 optimizations like field reordering, but it can be fixed with the
3721 `#[repr(...)]` attribute. In either case, fields may be given in any order in
3722 a corresponding struct *expression*; the resulting `struct` value will always
3723 have the same memory layout.
3725 The fields of a `struct` may be qualified by [visibility
3726 modifiers](#re-exporting-and-visibility), to allow access to data in a
3727 structure outside a module.
3729 A _tuple struct_ type is just like a structure type, except that the fields are
3732 A _unit-like struct_ type is like a structure type, except that it has no
3733 fields. The one value constructed by the associated [structure
3734 expression](#structure-expressions) is the only value that inhabits such a
3737 ### Enumerated types
3739 An *enumerated type* is a nominal, heterogeneous disjoint union type, denoted
3740 by the name of an [`enum` item](#enumerations). [^enumtype]
3742 [^enumtype]: The `enum` type is analogous to a `data` constructor declaration in
3743 ML, or a *pick ADT* in Limbo.
3745 An [`enum` item](#enumerations) declares both the type and a number of *variant
3746 constructors*, each of which is independently named and takes an optional tuple
3749 New instances of an `enum` can be constructed by calling one of the variant
3750 constructors, in a [call expression](#call-expressions).
3752 Any `enum` value consumes as much memory as the largest variant constructor for
3753 its corresponding `enum` type.
3755 Enum types cannot be denoted *structurally* as types, but must be denoted by
3756 named reference to an [`enum` item](#enumerations).
3760 Nominal types — [enumerations](#enumerated-types) and
3761 [structures](#structure-types) — may be recursive. That is, each `enum`
3762 constructor or `struct` field may refer, directly or indirectly, to the
3763 enclosing `enum` or `struct` type itself. Such recursion has restrictions:
3765 * Recursive types must include a nominal type in the recursion
3766 (not mere [type definitions](#type-definitions),
3767 or other structural types such as [arrays](#array,-and-slice-types) or [tuples](#tuple-types)).
3768 * A recursive `enum` item must have at least one non-recursive constructor
3769 (in order to give the recursion a basis case).
3770 * The size of a recursive type must be finite;
3771 in other words the recursive fields of the type must be [pointer types](#pointer-types).
3772 * Recursive type definitions can cross module boundaries, but not module *visibility* boundaries,
3773 or crate boundaries (in order to simplify the module system and type checker).
3775 An example of a *recursive* type and its use:
3780 Cons(T, Box<List<T>>)
3783 let a: List<int> = List::Cons(7, box List::Cons(13, box List::Nil));
3788 All pointers in Rust are explicit first-class values. They can be copied,
3789 stored into data structures, and returned from functions. There are two
3790 varieties of pointer in Rust:
3793 : These point to memory _owned by some other value_.
3794 A reference type is written `&type` for some lifetime-variable `f`,
3795 or just `&'a type` when you need an explicit lifetime.
3796 Copying a reference is a "shallow" operation:
3797 it involves only copying the pointer itself.
3798 Releasing a reference typically has no effect on the value it points to,
3799 with the exception of temporary values, which are released when the last
3800 reference to them is released.
3802 * Raw pointers (`*`)
3803 : Raw pointers are pointers without safety or liveness guarantees.
3804 Raw pointers are written as `*const T` or `*mut T`,
3805 for example `*const int` means a raw pointer to an integer.
3806 Copying or dropping a raw pointer has no effect on the lifecycle of any
3807 other value. Dereferencing a raw pointer or converting it to any other
3808 pointer type is an [`unsafe` operation](#unsafe-functions).
3809 Raw pointers are generally discouraged in Rust code;
3810 they exist to support interoperability with foreign code,
3811 and writing performance-critical or low-level functions.
3813 The standard library contains additional 'smart pointer' types beyond references
3818 The function type constructor `fn` forms new function types. A function type
3819 consists of a possibly-empty set of function-type modifiers (such as `unsafe`
3820 or `extern`), a sequence of input types and an output type.
3822 An example of a `fn` type:
3825 fn add(x: int, y: int) -> int {
3829 let mut x = add(5,7);
3831 type Binop<'a> = |int,int|: 'a -> int;
3832 let bo: Binop = add;
3838 ```{.ebnf .notation}
3839 closure_type := [ 'unsafe' ] [ '<' lifetime-list '>' ] '|' arg-list '|'
3840 [ ':' bound-list ] [ '->' type ]
3841 lifetime-list := lifetime | lifetime ',' lifetime-list
3842 arg-list := ident ':' type | ident ':' type ',' arg-list
3843 bound-list := bound | bound '+' bound-list
3844 bound := path | lifetime
3847 The type of a closure mapping an input of type `A` to an output of type `B` is
3848 `|A| -> B`. A closure with no arguments or return values has type `||`.
3850 An example of creating and calling a closure:
3853 let captured_var = 10i;
3855 let closure_no_args = || println!("captured_var={}", captured_var);
3857 let closure_args = |arg: int| -> int {
3858 println!("captured_var={}, arg={}", captured_var, arg);
3859 arg // Note lack of semicolon after 'arg'
3862 fn call_closure(c1: ||, c2: |int| -> int) {
3867 call_closure(closure_no_args, closure_args);
3873 Every trait item (see [traits](#traits)) defines a type with the same name as
3874 the trait. This type is called the _object type_ of the trait. Object types
3875 permit "late binding" of methods, dispatched using _virtual method tables_
3876 ("vtables"). Whereas most calls to trait methods are "early bound" (statically
3877 resolved) to specific implementations at compile time, a call to a method on an
3878 object type is only resolved to a vtable entry at compile time. The actual
3879 implementation for each vtable entry can vary on an object-by-object basis.
3881 Given a pointer-typed expression `E` of type `&T` or `Box<T>`, where `T`
3882 implements trait `R`, casting `E` to the corresponding pointer type `&R` or
3883 `Box<R>` results in a value of the _object type_ `R`. This result is
3884 represented as a pair of pointers: the vtable pointer for the `T`
3885 implementation of `R`, and the pointer value of `E`.
3887 An example of an object type:
3891 fn stringify(&self) -> String;
3894 impl Printable for int {
3895 fn stringify(&self) -> String { self.to_string() }
3898 fn print(a: Box<Printable>) {
3899 println!("{}", a.stringify());
3903 print(box 10i as Box<Printable>);
3907 In this example, the trait `Printable` occurs as an object type in both the
3908 type signature of `print`, and the cast expression in `main`.
3912 Within the body of an item that has type parameter declarations, the names of
3913 its type parameters are types:
3916 fn map<A: Clone, B: Clone>(f: |A| -> B, xs: &[A]) -> Vec<B> {
3920 let first: B = f(xs[0].clone());
3921 let mut rest: Vec<B> = map(f, xs.slice(1, xs.len()));
3922 rest.insert(0, first);
3927 Here, `first` has type `B`, referring to `map`'s `B` type parameter; and `rest`
3928 has type `Vec<B>`, a vector type with element type `B`.
3932 The special type `self` has a meaning within methods inside an impl item. It
3933 refers to the type of the implicit `self` argument. For example, in:
3937 fn make_string(&self) -> String;
3940 impl Printable for String {
3941 fn make_string(&self) -> String {
3947 `self` refers to the value of type `String` that is the receiver for a call to
3948 the method `make_string`.
3952 Types in Rust are categorized into kinds, based on various properties of the
3953 components of the type. The kinds are:
3956 : Types of this kind can be safely sent between threads.
3957 This kind includes scalars, boxes, procs, and
3958 structural types containing only other owned types.
3959 All `Send` types are `'static`.
3961 : Types of this kind consist of "Plain Old Data"
3962 which can be copied by simply moving bits.
3963 All values of this kind can be implicitly copied.
3964 This kind includes scalars and immutable references,
3965 as well as structural types containing other `Copy` types.
3967 : Types of this kind do not contain any references (except for
3968 references with the `static` lifetime, which are allowed).
3969 This can be a useful guarantee for code
3970 that breaks borrowing assumptions
3971 using [`unsafe` operations](#unsafe-functions).
3973 : This is not strictly a kind,
3974 but its presence interacts with kinds:
3975 the `Drop` trait provides a single method `drop`
3976 that takes no parameters,
3977 and is run when values of the type are dropped.
3978 Such a method is called a "destructor",
3979 and are always executed in "top-down" order:
3980 a value is completely destroyed
3981 before any of the values it owns run their destructors.
3982 Only `Send` types can implement `Drop`.
3985 : Types with destructors, closure environments,
3986 and various other _non-first-class_ types,
3987 are not copyable at all.
3988 Such types can usually only be accessed through pointers,
3989 or in some cases, moved between mutable locations.
3991 Kinds can be supplied as _bounds_ on type parameters, like traits, in which
3992 case the parameter is constrained to types satisfying that kind.
3994 By default, type parameters do not carry any assumed kind-bounds at all. When
3995 instantiating a type parameter, the kind bounds on the parameter are checked to
3996 be the same or narrower than the kind of the type that it is instantiated with.
3998 Sending operations are not part of the Rust language, but are implemented in
3999 the library. Generic functions that send values bound the kind of these values
4002 # Memory and concurrency models
4004 Rust has a memory model centered around concurrently-executing _threads_. Thus
4005 its memory model and its concurrency model are best discussed simultaneously,
4006 as parts of each only make sense when considered from the perspective of the
4009 When reading about the memory model, keep in mind that it is partitioned in
4010 order to support threads; and when reading about threads, keep in mind that their
4011 isolation and communication mechanisms are only possible due to the ownership
4012 and lifetime semantics of the memory model.
4016 A Rust program's memory consists of a static set of *items*, a set of
4017 [threads](#threads) each with its own *stack*, and a *heap*. Immutable portions of
4018 the heap may be shared between threads, mutable portions may not.
4020 Allocations in the stack consist of *slots*, and allocations in the heap
4023 ### Memory allocation and lifetime
4025 The _items_ of a program are those functions, modules and types that have their
4026 value calculated at compile-time and stored uniquely in the memory image of the
4027 rust process. Items are neither dynamically allocated nor freed.
4029 A thread's _stack_ consists of activation frames automatically allocated on entry
4030 to each function as the thread executes. A stack allocation is reclaimed when
4031 control leaves the frame containing it.
4033 The _heap_ is a general term that describes boxes. The lifetime of an
4034 allocation in the heap depends on the lifetime of the box values pointing to
4035 it. Since box values may themselves be passed in and out of frames, or stored
4036 in the heap, heap allocations may outlive the frame they are allocated within.
4038 ### Memory ownership
4040 A thread owns all memory it can *safely* reach through local variables, as well
4041 as boxes and references.
4043 When a thread sends a value that has the `Send` trait to another thread, it loses
4044 ownership of the value sent and can no longer refer to it. This is statically
4045 guaranteed by the combined use of "move semantics", and the compiler-checked
4046 _meaning_ of the `Send` trait: it is only instantiated for (transitively)
4047 sendable kinds of data constructor and pointers, never including references.
4049 When a stack frame is exited, its local allocations are all released, and its
4050 references to boxes are dropped.
4052 When a thread finishes, its stack is necessarily empty and it therefore has no
4053 references to any boxes; the remainder of its heap is immediately freed.
4057 A thread's stack contains slots.
4059 A _slot_ is a component of a stack frame, either a function parameter, a
4060 [temporary](#lvalues,-rvalues-and-temporaries), or a local variable.
4062 A _local variable_ (or *stack-local* allocation) holds a value directly,
4063 allocated within the stack's memory. The value is a part of the stack frame.
4065 Local variables are immutable unless declared otherwise like: `let mut x = ...`.
4067 Function parameters are immutable unless declared with `mut`. The `mut` keyword
4068 applies only to the following parameter (so `|mut x, y|` and `fn f(mut x:
4069 Box<int>, y: Box<int>)` declare one mutable variable `x` and one immutable
4072 Methods that take either `self` or `Box<Self>` can optionally place them in a
4073 mutable slot by prefixing them with `mut` (similar to regular arguments):
4077 fn change(mut self) -> Self;
4078 fn modify(mut self: Box<Self>) -> Box<Self>;
4082 Local variables are not initialized when allocated; the entire frame worth of
4083 local variables are allocated at once, on frame-entry, in an uninitialized
4084 state. Subsequent statements within a function may or may not initialize the
4085 local variables. Local variables can be used only after they have been
4086 initialized; this is enforced by the compiler.
4090 A _box_ is a reference to a heap allocation holding another value, which is
4091 constructed by the prefix operator `box`. When the standard library is in use,
4092 the type of a box is `std::owned::Box<T>`.
4094 An example of a box type and value:
4097 let x: Box<int> = box 10;
4100 Box values exist in 1:1 correspondence with their heap allocation, copying a
4101 box value makes a shallow copy of the pointer. Rust will consider a shallow
4102 copy of a box to move ownership of the value. After a value has been moved,
4103 the source location cannot be used unless it is reinitialized.
4106 let x: Box<int> = box 10;
4108 // attempting to use `x` will result in an error here
4113 Rust's primary concurrency mechanism is called a **thread**.
4115 ### Communication between threads
4117 Rust threads are isolated and generally unable to interfere with one another's
4118 memory directly, except through [`unsafe` code](#unsafe-functions). All
4119 contact between threads is mediated by safe forms of ownership transfer, and data
4120 races on memory are prohibited by the type system.
4122 When you wish to send data between threads, the values are restricted to the
4123 [`Send` type-kind](#type-kinds). Restricting communication interfaces to this
4124 kind ensures that no references move between threads. Thus access to an entire
4125 data structure can be mediated through its owning "root" value; no further
4126 locking or copying is required to avoid data races within the substructure of
4131 The _lifecycle_ of a threads consists of a finite set of states and events that
4132 cause transitions between the states. The lifecycle states of a thread are:
4139 A thread begins its lifecycle — once it has been spawned — in the
4140 *running* state. In this state it executes the statements of its entry
4141 function, and any functions called by the entry function.
4143 A thread may transition from the *running* state to the *blocked* state any time
4144 it makes a blocking communication call. When the call can be completed —
4145 when a message arrives at a sender, or a buffer opens to receive a message
4146 — then the blocked thread will unblock and transition back to *running*.
4148 A thread may transition to the *panicked* state at any time, due being killed by
4149 some external event or internally, from the evaluation of a `panic!()` macro.
4150 Once *panicking*, a thread unwinds its stack and transitions to the *dead* state.
4151 Unwinding the stack of a thread is done by the thread itself, on its own control
4152 stack. If a value with a destructor is freed during unwinding, the code for the
4153 destructor is run, also on the thread's control stack. Running the destructor
4154 code causes a temporary transition to a *running* state, and allows the
4155 destructor code to cause any subsequent state transitions. The original thread
4156 of unwinding and panicking thereby may suspend temporarily, and may involve
4157 (recursive) unwinding of the stack of a failed destructor. Nonetheless, the
4158 outermost unwinding activity will continue until the stack is unwound and the
4159 thread transitions to the *dead* state. There is no way to "recover" from thread
4160 panics. Once a thread has temporarily suspended its unwinding in the *panicking*
4161 state, a panic occurring from within this destructor results in *hard* panic.
4162 A hard panic currently results in the process aborting.
4164 A thread in the *dead* state cannot transition to other states; it exists only to
4165 have its termination status inspected by other threads, and/or to await
4166 reclamation when the last reference to it drops.
4168 # Runtime services, linkage and debugging
4170 The Rust _runtime_ is a relatively compact collection of Rust code that
4171 provides fundamental services and datatypes to all Rust threads at run-time. It
4172 is smaller and simpler than many modern language runtimes. It is tightly
4173 integrated into the language's execution model of memory, threads, communication
4176 ### Memory allocation
4178 The runtime memory-management system is based on a _service-provider
4179 interface_, through which the runtime requests blocks of memory from its
4180 environment and releases them back to its environment when they are no longer
4181 needed. The default implementation of the service-provider interface consists
4182 of the C runtime functions `malloc` and `free`.
4184 The runtime memory-management system, in turn, supplies Rust threads with
4185 facilities for allocating releasing stacks, as well as allocating and freeing
4190 The runtime provides C and Rust code to assist with various built-in types,
4191 such as arrays, strings, and the low level communication system (ports,
4194 Support for other built-in types such as simple types, tuples and enums is
4195 open-coded by the Rust compiler.
4197 ### Thread scheduling and communication
4199 The runtime provides code to manage inter-thread communication. This includes
4200 the system of thread-lifecycle state transitions depending on the contents of
4201 queues, as well as code to copy values between queues and their recipients and
4202 to serialize values for transmission over operating-system inter-process
4203 communication facilities.
4207 The Rust compiler supports various methods to link crates together both
4208 statically and dynamically. This section will explore the various methods to
4209 link Rust crates together, and more information about native libraries can be
4210 found in the [ffi guide][ffi].
4212 In one session of compilation, the compiler can generate multiple artifacts
4213 through the usage of either command line flags or the `crate_type` attribute.
4214 If one or more command line flag is specified, all `crate_type` attributes will
4215 be ignored in favor of only building the artifacts specified by command line.
4217 * `--crate-type=bin`, `#[crate_type = "bin"]` - A runnable executable will be
4218 produced. This requires that there is a `main` function in the crate which
4219 will be run when the program begins executing. This will link in all Rust and
4220 native dependencies, producing a distributable binary.
4222 * `--crate-type=lib`, `#[crate_type = "lib"]` - A Rust library will be produced.
4223 This is an ambiguous concept as to what exactly is produced because a library
4224 can manifest itself in several forms. The purpose of this generic `lib` option
4225 is to generate the "compiler recommended" style of library. The output library
4226 will always be usable by rustc, but the actual type of library may change from
4227 time-to-time. The remaining output types are all different flavors of
4228 libraries, and the `lib` type can be seen as an alias for one of them (but the
4229 actual one is compiler-defined).
4231 * `--crate-type=dylib`, `#[crate_type = "dylib"]` - A dynamic Rust library will
4232 be produced. This is different from the `lib` output type in that this forces
4233 dynamic library generation. The resulting dynamic library can be used as a
4234 dependency for other libraries and/or executables. This output type will
4235 create `*.so` files on linux, `*.dylib` files on osx, and `*.dll` files on
4238 * `--crate-type=staticlib`, `#[crate_type = "staticlib"]` - A static system
4239 library will be produced. This is different from other library outputs in that
4240 the Rust compiler will never attempt to link to `staticlib` outputs. The
4241 purpose of this output type is to create a static library containing all of
4242 the local crate's code along with all upstream dependencies. The static
4243 library is actually a `*.a` archive on linux and osx and a `*.lib` file on
4244 windows. This format is recommended for use in situations such as linking
4245 Rust code into an existing non-Rust application because it will not have
4246 dynamic dependencies on other Rust code.
4248 * `--crate-type=rlib`, `#[crate_type = "rlib"]` - A "Rust library" file will be
4249 produced. This is used as an intermediate artifact and can be thought of as a
4250 "static Rust library". These `rlib` files, unlike `staticlib` files, are
4251 interpreted by the Rust compiler in future linkage. This essentially means
4252 that `rustc` will look for metadata in `rlib` files like it looks for metadata
4253 in dynamic libraries. This form of output is used to produce statically linked
4254 executables as well as `staticlib` outputs.
4256 Note that these outputs are stackable in the sense that if multiple are
4257 specified, then the compiler will produce each form of output at once without
4258 having to recompile. However, this only applies for outputs specified by the
4259 same method. If only `crate_type` attributes are specified, then they will all
4260 be built, but if one or more `--crate-type` command line flag is specified,
4261 then only those outputs will be built.
4263 With all these different kinds of outputs, if crate A depends on crate B, then
4264 the compiler could find B in various different forms throughout the system. The
4265 only forms looked for by the compiler, however, are the `rlib` format and the
4266 dynamic library format. With these two options for a dependent library, the
4267 compiler must at some point make a choice between these two formats. With this
4268 in mind, the compiler follows these rules when determining what format of
4269 dependencies will be used:
4271 1. If a static library is being produced, all upstream dependencies are
4272 required to be available in `rlib` formats. This requirement stems from the
4273 reason that a dynamic library cannot be converted into a static format.
4275 Note that it is impossible to link in native dynamic dependencies to a static
4276 library, and in this case warnings will be printed about all unlinked native
4277 dynamic dependencies.
4279 2. If an `rlib` file is being produced, then there are no restrictions on what
4280 format the upstream dependencies are available in. It is simply required that
4281 all upstream dependencies be available for reading metadata from.
4283 The reason for this is that `rlib` files do not contain any of their upstream
4284 dependencies. It wouldn't be very efficient for all `rlib` files to contain a
4285 copy of `libstd.rlib`!
4287 3. If an executable is being produced and the `-C prefer-dynamic` flag is not
4288 specified, then dependencies are first attempted to be found in the `rlib`
4289 format. If some dependencies are not available in an rlib format, then
4290 dynamic linking is attempted (see below).
4292 4. If a dynamic library or an executable that is being dynamically linked is
4293 being produced, then the compiler will attempt to reconcile the available
4294 dependencies in either the rlib or dylib format to create a final product.
4296 A major goal of the compiler is to ensure that a library never appears more
4297 than once in any artifact. For example, if dynamic libraries B and C were
4298 each statically linked to library A, then a crate could not link to B and C
4299 together because there would be two copies of A. The compiler allows mixing
4300 the rlib and dylib formats, but this restriction must be satisfied.
4302 The compiler currently implements no method of hinting what format a library
4303 should be linked with. When dynamically linking, the compiler will attempt to
4304 maximize dynamic dependencies while still allowing some dependencies to be
4305 linked in via an rlib.
4307 For most situations, having all libraries available as a dylib is recommended
4308 if dynamically linking. For other situations, the compiler will emit a
4309 warning if it is unable to determine which formats to link each library with.
4311 In general, `--crate-type=bin` or `--crate-type=lib` should be sufficient for
4312 all compilation needs, and the other options are just available if more
4313 fine-grained control is desired over the output format of a Rust crate.
4315 # Appendix: Rationales and design tradeoffs
4319 # Appendix: Influences and further references
4323 > The essential problem that must be solved in making a fault-tolerant
4324 > software system is therefore that of fault-isolation. Different programmers
4325 > will write different modules, some modules will be correct, others will have
4326 > errors. We do not want the errors in one module to adversely affect the
4327 > behaviour of a module which does not have any errors.
4329 > — Joe Armstrong
4331 > In our approach, all data is private to some process, and processes can
4332 > only communicate through communications channels. *Security*, as used
4333 > in this paper, is the property which guarantees that processes in a system
4334 > cannot affect each other except by explicit communication.
4336 > When security is absent, nothing which can be proven about a single module
4337 > in isolation can be guaranteed to hold when that module is embedded in a
4340 > — Robert Strom and Shaula Yemini
4342 > Concurrent and applicative programming complement each other. The
4343 > ability to send messages on channels provides I/O without side effects,
4344 > while the avoidance of shared data helps keep concurrent processes from
4349 Rust is not a particularly original language. It may however appear unusual by
4350 contemporary standards, as its design elements are drawn from a number of
4351 "historical" languages that have, with a few exceptions, fallen out of favour.
4352 Five prominent lineages contribute the most, though their influences have come
4353 and gone during the course of Rust's development:
4355 * The NIL (1981) and Hermes (1990) family. These languages were developed by
4356 Robert Strom, Shaula Yemini, David Bacon and others in their group at IBM
4357 Watson Research Center (Yorktown Heights, NY, USA).
4359 * The Erlang (1987) language, developed by Joe Armstrong, Robert Virding, Claes
4360 Wikström, Mike Williams and others in their group at the Ericsson Computer
4361 Science Laboratory (Älvsjö, Stockholm, Sweden) .
4363 * The Sather (1990) language, developed by Stephen Omohundro, Chu-Cheow Lim,
4364 Heinz Schmidt and others in their group at The International Computer
4365 Science Institute of the University of California, Berkeley (Berkeley, CA,
4368 * The Newsqueak (1988), Alef (1995), and Limbo (1996) family. These
4369 languages were developed by Rob Pike, Phil Winterbottom, Sean Dorward and
4370 others in their group at Bell Labs Computing Sciences Research Center
4371 (Murray Hill, NJ, USA).
4373 * The Napier (1985) and Napier88 (1988) family. These languages were
4374 developed by Malcolm Atkinson, Ron Morrison and others in their group at
4375 the University of St. Andrews (St. Andrews, Fife, UK).
4377 Additional specific influences can be seen from the following languages:
4379 * The structural algebraic types and compilation manager of SML.
4380 * The attribute and assembly systems of C#.
4381 * The references and deterministic destructor system of C++.
4382 * The memory region systems of the ML Kit and Cyclone.
4383 * The typeclass system of Haskell.
4384 * The lexical identifier rule of Python.
4385 * The block syntax of Ruby.
4387 [ffi]: guide-ffi.html
4388 [plugin]: guide-plugin.html