1 % The Rust Reference Manual
5 This document is the reference manual 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 a tutorial introduction to the
17 language. Background familiarity with the language is assumed. A separate
18 [tutorial] document is available to help acquire such background familiarity.
20 This document also does not serve as a reference to the [standard] or [extra]
21 libraries included in the language distribution. Those libraries are
22 documented separately by extracting documentation attributes from their
25 [tutorial]: tutorial.html
26 [standard]: std/index.html
27 [extra]: extra/index.html
31 Rust is a work in progress. The language continues to evolve as the design
32 shifts and is fleshed out in working code. Certain parts work, certain parts
33 do not, certain parts will be removed or changed.
35 This manual is a snapshot written in the present tense. All features described
36 exist in working code unless otherwise noted, but some are quite primitive or
37 remain to be further modified by planned work. Some may be temporary. It is a
38 *draft*, and we ask that you not take anything you read here as final.
40 If you have suggestions to make, please try to focus them on *reductions* to
41 the language: possible features that can be combined or omitted. We aim to
42 keep the size and complexity of the language under control.
44 > **Note:** The grammar for Rust given in this document is rough and
45 > very incomplete; only a modest number of sections have accompanying grammar
46 > rules. Formalizing the grammar accepted by the Rust parser is ongoing work,
47 > but future versions of this document will contain a complete
48 > grammar. Moreover, we hope that this grammar will be extracted and verified
49 > as LL(1) by an automated grammar-analysis tool, and further tested against the
50 > Rust sources. Preliminary versions of this automation exist, but are not yet
55 Rust's grammar is defined over Unicode codepoints, each conventionally
56 denoted `U+XXXX`, for 4 or more hexadecimal digits `X`. _Most_ of Rust's
57 grammar is confined to the ASCII range of Unicode, and is described in this
58 document by a dialect of Extended Backus-Naur Form (EBNF), specifically a
59 dialect of EBNF supported by common automated LL(k) parsing tools such as
60 `llgen`, rather than the dialect given in ISO 14977. The dialect can be
61 defined self-referentially as follows:
63 ~~~~ {.notrust .ebnf .notation}
65 rule : nonterminal ':' productionrule ';' ;
66 productionrule : production [ '|' production ] * ;
68 term : element repeats ;
69 element : LITERAL | IDENTIFIER | '[' productionrule ']' ;
70 repeats : [ '*' | '+' ] NUMBER ? | NUMBER ? | '?' ;
75 - Whitespace in the grammar is ignored.
76 - Square brackets are used to group rules.
77 - `LITERAL` is a single printable ASCII character, or an escaped hexadecimal
78 ASCII code of the form `\xQQ`, in single quotes, denoting the corresponding
79 Unicode codepoint `U+00QQ`.
80 - `IDENTIFIER` is a nonempty string of ASCII letters and underscores.
81 - The `repeat` forms apply to the adjacent `element`, and are as follows:
82 - `?` means zero or one repetition
83 - `*` means zero or more repetitions
84 - `+` means one or more repetitions
85 - NUMBER trailing a repeat symbol gives a maximum repetition count
86 - NUMBER on its own gives an exact repetition count
88 This EBNF dialect should hopefully be familiar to many readers.
90 ## Unicode productions
92 A few productions in Rust's grammar permit Unicode codepoints outside the ASCII range.
93 We define these productions in terms of character properties specified in the Unicode standard,
94 rather than in terms of ASCII-range codepoints.
95 The section [Special Unicode Productions](#special-unicode-productions) lists these productions.
97 ## String table productions
99 Some rules in the grammar -- notably [unary
100 operators](#unary-operator-expressions), [binary
101 operators](#binary-operator-expressions), and [keywords](#keywords) --
102 are given in a simplified form: as a listing of a table of unquoted,
103 printable whitespace-separated strings. These cases form a subset of
104 the rules regarding the [token](#tokens) rule, and are assumed to be
105 the result of a lexical-analysis phase feeding the parser, driven by a
106 DFA, operating over the disjunction of all such string table entries.
108 When such a string enclosed in double-quotes (`"`) occurs inside the
109 grammar, it is an implicit reference to a single member of such a string table
110 production. See [tokens](#tokens) for more information.
116 Rust input is interpreted as a sequence of Unicode codepoints encoded in UTF-8,
117 normalized to Unicode normalization form NFKC.
118 Most Rust grammar rules are defined in terms of printable ASCII-range codepoints,
119 but a small number are defined in terms of Unicode properties or explicit codepoint lists.
120 ^[Substitute definitions for the special Unicode productions are provided to the grammar verifier, restricted to ASCII range, when verifying the grammar in this document.]
122 ## Special Unicode Productions
124 The following productions in the Rust grammar are defined in terms of Unicode properties:
125 `ident`, `non_null`, `non_star`, `non_eol`, `non_slash_or_star`, `non_single_quote` and `non_double_quote`.
129 The `ident` production is any nonempty Unicode string of the following form:
131 - The first character has property `XID_start`
132 - The remaining characters have property `XID_continue`
134 that does _not_ occur in the set of [keywords](#keywords).
136 Note: `XID_start` and `XID_continue` as character properties cover the
137 character ranges used to form the more familiar C and Java language-family
140 ### Delimiter-restricted productions
142 Some productions are defined by exclusion of particular Unicode characters:
144 - `non_null` is any single Unicode character aside from `U+0000` (null)
145 - `non_eol` is `non_null` restricted to exclude `U+000A` (`'\n'`)
146 - `non_star` is `non_null` restricted to exclude `U+002A` (`*`)
147 - `non_slash_or_star` is `non_null` restricted to exclude `U+002F` (`/`) and `U+002A` (`*`)
148 - `non_single_quote` is `non_null` restricted to exclude `U+0027` (`'`)
149 - `non_double_quote` is `non_null` restricted to exclude `U+0022` (`"`)
153 ~~~~ {.notrust .ebnf .gram}
154 comment : block_comment | line_comment ;
155 block_comment : "/*" block_comment_body * '*' + '/' ;
156 block_comment_body : [block_comment | character] * ;
157 line_comment : "//" non_eol * ;
160 Comments in Rust code follow the general C++ style of line and block-comment forms,
161 with no nesting of block-comment delimiters.
163 Line comments beginning with exactly _three_ slashes (`///`), and block
164 comments beginning with a exactly one repeated asterisk in the block-open
165 sequence (`/**`), are interpreted as a special syntax for `doc`
166 [attributes](#attributes). That is, they are equivalent to writing
167 `#[doc="..."]` around the body of the comment (this includes the comment
168 characters themselves, ie `/// Foo` turns into `#[doc="/// Foo"]`).
170 Non-doc comments are interpreted as a form of whitespace.
174 ~~~~ {.notrust .ebnf .gram}
175 whitespace_char : '\x20' | '\x09' | '\x0a' | '\x0d' ;
176 whitespace : [ whitespace_char | comment ] + ;
179 The `whitespace_char` production is any nonempty Unicode string consisting of any
180 of the following Unicode characters: `U+0020` (space, `' '`), `U+0009` (tab,
181 `'\t'`), `U+000A` (LF, `'\n'`), `U+000D` (CR, `'\r'`).
183 Rust is a "free-form" language, meaning that all forms of whitespace serve
184 only to separate _tokens_ in the grammar, and have no semantic significance.
186 A Rust program has identical meaning if each whitespace element is replaced
187 with any other legal whitespace element, such as a single space character.
191 ~~~~ {.notrust .ebnf .gram}
192 simple_token : keyword | unop | binop ;
193 token : simple_token | ident | literal | symbol | whitespace token ;
196 Tokens are primitive productions in the grammar defined by regular
197 (non-recursive) languages. "Simple" tokens are given in [string table
198 production](#string-table-productions) form, and occur in the rest of the
199 grammar as double-quoted strings. Other tokens have exact rules given.
203 The keywords are the following strings:
205 ~~~~ {.notrust .keyword}
216 self static struct super
222 Each of these keywords has special meaning in its grammar,
223 and all of them are excluded from the `ident` rule.
227 A literal is an expression consisting of a single token, rather than a
228 sequence of tokens, that immediately and directly denotes the value it
229 evaluates to, rather than referring to it by name or some other evaluation
230 rule. A literal is a form of constant expression, so is evaluated (primarily)
233 ~~~~ {.notrust .ebnf .gram}
234 literal : string_lit | char_lit | num_lit ;
237 #### Character and string literals
239 ~~~~ {.notrust .ebnf .gram}
240 char_lit : '\x27' char_body '\x27' ;
241 string_lit : '"' string_body * '"' | 'r' raw_string ;
243 char_body : non_single_quote
244 | '\x5c' [ '\x27' | common_escape ] ;
246 string_body : non_double_quote
247 | '\x5c' [ '\x22' | common_escape ] ;
248 raw_string : '"' raw_string_body '"' | '#' raw_string '#' ;
250 common_escape : '\x5c'
251 | 'n' | 'r' | 't' | '0'
256 hex_digit : 'a' | 'b' | 'c' | 'd' | 'e' | 'f'
257 | 'A' | 'B' | 'C' | 'D' | 'E' | 'F'
259 oct_digit : '0' | '1' | '2' | '3' | '4' | '5' | '6' | '7' ;
260 dec_digit : '0' | nonzero_dec ;
261 nonzero_dec: '1' | '2' | '3' | '4'
262 | '5' | '6' | '7' | '8' | '9' ;
265 A _character literal_ is a single Unicode character enclosed within two
266 `U+0027` (single-quote) characters, with the exception of `U+0027` itself,
267 which must be _escaped_ by a preceding U+005C character (`\`).
269 A _string literal_ is a sequence of any Unicode characters enclosed within
270 two `U+0022` (double-quote) characters, with the exception of `U+0022`
271 itself, which must be _escaped_ by a preceding `U+005C` character (`\`),
272 or a _raw string literal_.
274 Some additional _escapes_ are available in either character or non-raw string
275 literals. An escape starts with a `U+005C` (`\`) and continues with one of
278 * An _8-bit codepoint escape_ escape starts with `U+0078` (`x`) and is
279 followed by exactly two _hex digits_. It denotes the Unicode codepoint
280 equal to the provided hex value.
281 * A _16-bit codepoint escape_ starts with `U+0075` (`u`) and is followed
282 by exactly four _hex digits_. It denotes the Unicode codepoint equal to
283 the provided hex value.
284 * A _32-bit codepoint escape_ starts with `U+0055` (`U`) and is followed
285 by exactly eight _hex digits_. It denotes the Unicode codepoint equal to
286 the provided hex value.
287 * A _whitespace escape_ is one of the characters `U+006E` (`n`), `U+0072`
288 (`r`), or `U+0074` (`t`), denoting the unicode values `U+000A` (LF),
289 `U+000D` (CR) or `U+0009` (HT) respectively.
290 * The _backslash escape_ is the character `U+005C` (`\`) which must be
291 escaped in order to denote *itself*.
293 Raw string literals do not process any escapes. They start with the character
294 `U+0072` (`r`), followed zero or more of the character `U+0023` (`#`) and a
295 `U+0022` (double-quote) character. The _raw string body_ is not defined in the
296 EBNF grammar above: it can contain any sequence of Unicode characters and is
297 terminated only by another `U+0022` (double-quote) character, followed by the
298 same number of `U+0023` (`#`) characters that preceeded the opening `U+0022`
299 (double-quote) character.
301 All Unicode characters contained in the raw string body represent themselves,
302 the characters `U+0022` (double-quote) (except when followed by at least as
303 many `U+0023` (`#`) characters as were used to start the raw string literal) or
304 `U+005C` (`\`) do not have any special meaning.
306 Examples for string literals:
309 "foo"; r"foo"; // foo
310 "\"foo\""; r#""foo""#; // "foo"
313 r##"foo #"# bar"##; // foo #"# bar
315 "\x52"; "R"; r"R"; // R
316 "\\x52"; r"\x52"; // \x52
321 ~~~~ {.notrust .ebnf .gram}
322 num_lit : nonzero_dec [ dec_digit | '_' ] * num_suffix ?
323 | '0' [ [ dec_digit | '_' ] * num_suffix ?
324 | 'b' [ '1' | '0' | '_' ] + int_suffix ?
325 | 'o' [ oct_digit | '_' ] + int_suffix ?
326 | 'x' [ hex_digit | '_' ] + int_suffix ? ] ;
328 num_suffix : int_suffix | float_suffix ;
330 int_suffix : 'u' int_suffix_size ?
331 | 'i' int_suffix_size ? ;
332 int_suffix_size : [ '8' | '1' '6' | '3' '2' | '6' '4' ] ;
334 float_suffix : [ exponent | '.' dec_lit exponent ? ] ? float_suffix_ty ? ;
335 float_suffix_ty : 'f' [ '3' '2' | '6' '4' ] ;
336 exponent : ['E' | 'e'] ['-' | '+' ] ? dec_lit ;
337 dec_lit : [ dec_digit | '_' ] + ;
340 A _number literal_ is either an _integer literal_ or a _floating-point
341 literal_. The grammar for recognizing the two kinds of literals is mixed,
342 as they are differentiated by suffixes.
344 ##### Integer literals
346 An _integer literal_ has one of four forms:
348 * A _decimal literal_ starts with a *decimal digit* and continues with any
349 mixture of *decimal digits* and _underscores_.
350 * A _hex literal_ starts with the character sequence `U+0030` `U+0078`
351 (`0x`) and continues as any mixture hex digits and underscores.
352 * An _octal literal_ starts with the character sequence `U+0030` `U+006F`
353 (`0o`) and continues as any mixture octal digits and underscores.
354 * A _binary literal_ starts with the character sequence `U+0030` `U+0062`
355 (`0b`) and continues as any mixture binary digits and underscores.
357 An integer literal may be followed (immediately, without any spaces) by an
358 _integer suffix_, which changes the type of the literal. There are two kinds
359 of integer literal suffix:
361 * The `i` and `u` suffixes give the literal type `int` or `uint`,
363 * Each of the signed and unsigned machine types `u8`, `i8`,
364 `u16`, `i16`, `u32`, `i32`, `u64` and `i64`
365 give the literal the corresponding machine type.
367 The type of an _unsuffixed_ integer literal is determined by type inference.
368 If a integer type can be _uniquely_ determined from the surrounding program
369 context, the unsuffixed integer literal has that type. If the program context
370 underconstrains the type, the unsuffixed integer literal's type is `int`; if
371 the program context overconstrains the type, it is considered a static type
374 Examples of integer literals of various forms:
377 123; 0xff00; // type determined by program context
378 // defaults to int in absence of type
384 0o70_i16; // type i16
385 0b1111_1111_1001_0000_i32; // type i32
388 ##### Floating-point literals
390 A _floating-point literal_ has one of two forms:
392 * Two _decimal literals_ separated by a period
393 character `U+002E` (`.`), with an optional _exponent_ trailing after the
394 second decimal literal.
395 * A single _decimal literal_ followed by an _exponent_.
397 By default, a floating-point literal has a generic type, but will fall back to
398 `f64`. A floating-point literal may be followed (immediately, without any
399 spaces) by a _floating-point suffix_, which changes the type of the literal.
400 There are two floating-point suffixes: `f32`, and `f64` (the 32-bit and 64-bit
401 floating point types).
403 Examples of floating-point literals of various forms:
409 12E+99_f64; // type f64
412 ##### Unit and boolean literals
414 The _unit value_, the only value of the type that has the same name, is written as `()`.
415 The two values of the boolean type are written `true` and `false`.
419 ~~~~ {.notrust .ebnf .gram}
421 | '#' | '[' | ']' | '(' | ')' | '{' | '}'
425 Symbols are a general class of printable [token](#tokens) that play structural
426 roles in a variety of grammar productions. They are catalogued here for
427 completeness as the set of remaining miscellaneous printable tokens that do not
428 otherwise appear as [unary operators](#unary-operator-expressions), [binary
429 operators](#binary-operator-expressions), or [keywords](#keywords).
434 ~~~~ {.notrust .ebnf .gram}
435 expr_path : [ "::" ] ident [ "::" expr_path_tail ] + ;
436 expr_path_tail : '<' type_expr [ ',' type_expr ] + '>'
439 type_path : ident [ type_path_tail ] + ;
440 type_path_tail : '<' type_expr [ ',' type_expr ] + '>'
444 A _path_ is a sequence of one or more path components _logically_ separated by
445 a namespace qualifier (`::`). If a path consists of only one component, it may
446 refer to either an [item](#items) or a [slot](#memory-slots) in a local
447 control scope. If a path has multiple components, it refers to an item.
449 Every item has a _canonical path_ within its crate, but the path naming an
450 item is only meaningful within a given crate. There is no global namespace
451 across crates; an item's canonical path merely identifies it within the crate.
453 Two examples of simple paths consisting of only identifier components:
460 Path components are usually [identifiers](#identifiers), but the trailing
461 component of a path may be an angle-bracket-enclosed list of type
462 arguments. In [expression](#expressions) context, the type argument list is
463 given after a final (`::`) namespace qualifier in order to disambiguate it
464 from a relational expression involving the less-than symbol (`<`). In type
465 expression context, the final namespace qualifier is omitted.
467 Two examples of paths with type arguments:
470 # struct HashMap<K, V>;
472 # fn id<T>(t: T) -> T { t }
473 type T = HashMap<int,~str>; // Type arguments used in a type expression
474 let x = id::<int>(10); // Type arguments used in a call expression
478 Paths can be denoted with various leading qualifiers to change the meaning of
481 * Paths starting with `::` are considered to be global paths where the
482 components of the path start being resolved from the crate root. Each
483 identifier in the path must resolve to an item.
491 ::a::foo(); // call a's foo function
497 * Paths starting with the keyword `super` begin resolution relative to the
498 parent module. Each further identifier must resolve to an item
506 super::a::foo(); // call a's foo function
512 * Paths starting with the keyword `self` begin resolution relative to the
513 current module. Each further identifier must resolve to an item.
525 A number of minor features of Rust are not central enough to have their own
526 syntax, and yet are not implementable as functions. Instead, they are given
527 names, and invoked through a consistent syntax: `name!(...)`. Examples
530 * `format!` : format data into a string
531 * `env!` : look up an environment variable's value at compile time
532 * `file!`: return the path to the file being compiled
533 * `stringify!` : pretty-print the Rust expression given as an argument
534 * `include!` : include the Rust expression in the given file
535 * `include_str!` : include the contents of the given file as a string
536 * `include_bin!` : include the contents of the given file as a binary blob
537 * `error!`, `warn!`, `info!`, `debug!` : provide diagnostic information.
539 All of the above extensions are expressions with values.
543 ~~~~ {.notrust .ebnf .gram}
544 expr_macro_rules : "macro_rules" '!' ident '(' macro_rule * ')' ;
545 macro_rule : '(' matcher * ')' "=>" '(' transcriber * ')' ';' ;
546 matcher : '(' matcher * ')' | '[' matcher * ']'
547 | '{' matcher * '}' | '$' ident ':' ident
548 | '$' '(' matcher * ')' sep_token? [ '*' | '+' ]
549 | non_special_token ;
550 transcriber : '(' transcriber * ')' | '[' transcriber * ']'
551 | '{' transcriber * '}' | '$' ident
552 | '$' '(' transcriber * ')' sep_token? [ '*' | '+' ]
553 | non_special_token ;
556 User-defined syntax extensions are called "macros",
557 and the `macro_rules` syntax extension defines them.
558 Currently, user-defined macros can expand to expressions, statements, or items.
560 (A `sep_token` is any token other than `*` and `+`.
561 A `non_special_token` is any token other than a delimiter or `$`.)
563 The macro expander looks up macro invocations by name,
564 and tries each macro rule in turn.
565 It transcribes the first successful match.
566 Matching and transcription are closely related to each other,
567 and we will describe them together.
571 The macro expander matches and transcribes every token that does not begin with a `$` literally, including delimiters.
572 For parsing reasons, delimiters must be balanced, but they are otherwise not special.
574 In the matcher, `$` _name_ `:` _designator_ matches the nonterminal in the
575 Rust syntax named by _designator_. Valid designators are `item`, `block`,
576 `stmt`, `pat`, `expr`, `ty` (type), `ident`, `path`, `matchers` (lhs of the `=>` in macro rules),
577 `tt` (rhs of the `=>` in macro rules). In the transcriber, the designator is already known, and so only
578 the name of a matched nonterminal comes after the dollar sign.
580 In both the matcher and transcriber, the Kleene star-like operator indicates repetition.
581 The Kleene star operator consists of `$` and parens, optionally followed by a separator token, followed by `*` or `+`.
582 `*` means zero or more repetitions, `+` means at least one repetition.
583 The parens are not matched or transcribed.
584 On the matcher side, a name is bound to _all_ of the names it
585 matches, in a structure that mimics the structure of the repetition
586 encountered on a successful match. The job of the transcriber is to sort that
589 The rules for transcription of these repetitions are called "Macro By Example".
590 Essentially, one "layer" of repetition is discharged at a time, and all of
591 them must be discharged by the time a name is transcribed. Therefore,
592 `( $( $i:ident ),* ) => ( $i )` is an invalid macro, but
593 `( $( $i:ident ),* ) => ( $( $i:ident ),* )` is acceptable (if trivial).
595 When Macro By Example encounters a repetition, it examines all of the `$`
596 _name_ s that occur in its body. At the "current layer", they all must repeat
597 the same number of times, so
598 ` ( $( $i:ident ),* ; $( $j:ident ),* ) => ( $( ($i,$j) ),* )` is valid if
599 given the argument `(a,b,c ; d,e,f)`, but not `(a,b,c ; d,e)`. The repetition
600 walks through the choices at that layer in lockstep, so the former input
601 transcribes to `( (a,d), (b,e), (c,f) )`.
603 Nested repetitions are allowed.
605 ### Parsing limitations
607 The parser used by the macro system is reasonably powerful, but the parsing of
608 Rust syntax is restricted in two ways:
610 1. The parser will always parse as much as possible. If it attempts to match
611 `$i:expr [ , ]` against `8 [ , ]`, it will attempt to parse `i` as an array
612 index operation and fail. Adding a separator can solve this problem.
613 2. The parser must have eliminated all ambiguity by the time it reaches a `$` _name_ `:` _designator_.
614 This requirement most often affects name-designator pairs when they occur at the beginning of, or immediately after, a `$(...)*`; requiring a distinctive token in front can solve the problem.
616 ## Syntax extensions useful for the macro author
618 * `log_syntax!` : print out the arguments at compile time
619 * `trace_macros!` : supply `true` or `false` to enable or disable macro expansion logging
620 * `stringify!` : turn the identifier argument into a string literal
621 * `concat!` : concatenates a comma-separated list of literals
622 * `concat_idents!` : create a new identifier by concatenating the arguments
624 # Crates and source files
626 Rust is a *compiled* language.
627 Its semantics obey a *phase distinction* between compile-time and run-time.
628 Those semantic rules that have a *static interpretation* govern the success or failure of compilation.
629 We refer to these rules as "static semantics".
630 Semantic rules called "dynamic semantics" govern the behavior of programs at run-time.
631 A program that fails to compile due to violation of a compile-time rule has no defined dynamic semantics; the compiler should halt with an error report, and produce no executable artifact.
633 The compilation model centres on artifacts called _crates_.
634 Each compilation processes a single crate in source form, and if successful, produces a single crate in binary form: either an executable or a library.^[A crate is somewhat
635 analogous to an *assembly* in the ECMA-335 CLI model, a *library* in the
636 SML/NJ Compilation Manager, a *unit* in the Owens and Flatt module system,
637 or a *configuration* in Mesa.]
639 A _crate_ is a unit of compilation and linking, as well as versioning, distribution and runtime loading.
640 A crate contains a _tree_ of nested [module](#modules) scopes.
641 The top level of this tree is a module that is anonymous (from the point of view of paths within the module) and any item within a crate has a canonical [module path](#paths) denoting its location within the crate's module tree.
643 The Rust compiler is always invoked with a single source file as input, and always produces a single output crate.
644 The processing of that source file may result in other source files being loaded as modules.
645 Source files have the extension `.rs`.
647 A Rust source file describes a module, the name and
648 location of which -- in the module tree of the current crate -- are defined
649 from outside the source file: either by an explicit `mod_item` in
650 a referencing source file, or by the name of the crate itself.
652 Each source file contains a sequence of zero or more `item` definitions,
653 and may optionally begin with any number of `attributes` that apply to the containing module.
654 Attributes on the anonymous crate module define important metadata that influences
655 the behavior of the compiler.
659 #![crate_id = "projx#2.5"]
661 // Additional metadata attributes
662 #![desc = "Project X"]
664 #![comment = "This is a comment on Project X."]
666 // Specify the output type
667 #![crate_type = "lib"]
670 #![warn(non_camel_case_types)]
673 A crate that contains a `main` function can be compiled to an executable.
674 If a `main` function is present, its return type must be [`unit`](#primitive-types) and it must take no arguments.
676 # Items and attributes
678 Crates contain [items](#items),
679 each of which may have some number of [attributes](#attributes) attached to it.
683 ~~~~ {.notrust .ebnf .gram}
684 item : mod_item | fn_item | type_item | struct_item | enum_item
685 | static_item | trait_item | impl_item | extern_block ;
688 An _item_ is a component of a crate; some module items can be defined in crate
689 files, but most are defined in source files. Items are organized within a
690 crate by a nested set of [modules](#modules). Every crate has a single
691 "outermost" anonymous module; all further items within the crate have
692 [paths](#paths) within the module tree of the crate.
694 Items are entirely determined at compile-time, generally remain fixed during
695 execution, and may reside in read-only memory.
697 There are several kinds of item:
699 * [modules](#modules)
700 * [functions](#functions)
701 * [type definitions](#type-definitions)
702 * [structures](#structures)
703 * [enumerations](#enumerations)
704 * [static items](#static-items)
706 * [implementations](#implementations)
708 Some items form an implicit scope for the declaration of sub-items. In other
709 words, within a function or module, declarations of items can (in many cases)
710 be mixed with the statements, control blocks, and similar artifacts that
711 otherwise compose the item body. The meaning of these scoped items is the same
712 as if the item was declared outside the scope -- it is still a static item --
713 except that the item's *path name* within the module namespace is qualified by
714 the name of the enclosing item, or is private to the enclosing item (in the
716 The grammar specifies the exact locations in which sub-item declarations may appear.
720 All items except modules may be *parameterized* by type. Type parameters are
721 given as a comma-separated list of identifiers enclosed in angle brackets
722 (`<...>`), after the name of the item and before its definition.
723 The type parameters of an item are considered "part of the name", not part of the type of the item.
724 A referencing [path](#paths) must (in principle) provide type arguments as a list of comma-separated types enclosed within angle brackets, in order to refer to the type-parameterized item.
725 In practice, the type-inference system can usually infer such argument types from context.
726 There are no general type-parametric types, only type-parametric items.
727 That is, Rust has no notion of type abstraction: there are no first-class "forall" types.
731 ~~~~ {.notrust .ebnf .gram}
732 mod_item : "mod" ident ( ';' | '{' mod '}' );
733 mod : [ view_item | item ] * ;
736 A module is a container for zero or more [view items](#view-items) and zero or
737 more [items](#items). The view items manage the visibility of the items
738 defined within the module, as well as the visibility of names from outside the
739 module when referenced from inside the module.
741 A _module item_ is a module, surrounded in braces, named, and prefixed with
742 the keyword `mod`. A module item introduces a new, named module into the tree
743 of modules making up a crate. Modules can nest arbitrarily.
745 An example of a module:
749 type Complex = (f64, f64);
750 fn sin(f: f64) -> f64 {
754 fn cos(f: f64) -> f64 {
758 fn tan(f: f64) -> f64 {
765 Modules and types share the same namespace.
766 Declaring a named type that has the same name as a module in scope is forbidden:
767 that is, a type definition, trait, struct, enumeration, or type parameter
768 can't shadow the name of a module in scope, or vice versa.
770 A module without a body is loaded from an external file, by default with the same
771 name as the module, plus the `.rs` extension.
772 When a nested submodule is loaded from an external file,
773 it is loaded from a subdirectory path that mirrors the module hierarchy.
776 // Load the `vec` module from `vec.rs`
780 // Load the `local_data` module from `task/local_data.rs`
785 The directories and files used for loading external file modules can be influenced
786 with the `path` attribute.
789 #[path = "task_files"]
791 // Load the `local_data` module from `task_files/tls.rs`
799 ~~~~ {.notrust .ebnf .gram}
800 view_item : extern_crate_decl | use_decl ;
803 A view item manages the namespace of a module.
804 View items do not define new items, but rather, simply change other items' visibility.
805 There are several kinds of view item:
807 * [`extern crate` declarations](#extern-crate-declarations)
808 * [`use` declarations](#use-declarations)
810 ##### Extern crate declarations
812 ~~~~ {.notrust .ebnf .gram}
813 extern_crate_decl : "extern" "crate" ident [ '(' link_attrs ')' ] ? [ '=' string_lit ] ? ;
814 link_attrs : link_attr [ ',' link_attrs ] + ;
815 link_attr : ident '=' literal ;
818 An _`extern crate` declaration_ specifies a dependency on an external crate.
819 The external crate is then bound into the declaring scope as the `ident` provided
820 in the `extern_crate_decl`.
822 The external crate is resolved to a specific `soname` at compile time, and a
823 runtime linkage requirement to that `soname` is passed to the linker for
824 loading at runtime. The `soname` is resolved at compile time by scanning the
825 compiler's library path and matching the optional `crateid` provided as a string literal
826 against the `crateid` attributes that were declared on the external crate when
827 it was compiled. If no `crateid` is provided, a default `name` attribute is
828 assumed, equal to the `ident` given in the `extern_crate_decl`.
830 Four examples of `extern crate` declarations:
835 extern crate std; // equivalent to: extern crate std = "std";
837 extern crate ruststd = "std"; // linking to 'std' under another name
839 extern crate foo = "some/where/rust-foo#foo:1.0"; // a full crate ID for external tools
842 ##### Use declarations
844 ~~~~ {.notrust .ebnf .gram}
845 use_decl : "pub" ? "use" ident [ '=' path
848 path_glob : ident [ "::" path_glob ] ?
850 | '{' ident [ ',' ident ] * '}' ;
853 A _use declaration_ creates one or more local name bindings synonymous
854 with some other [path](#paths).
855 Usually a `use` declaration is used to shorten the path required to refer to a
856 module item. These declarations may appear at the top of [modules](#modules) and
859 *Note*: Unlike in many languages,
860 `use` declarations in Rust do *not* declare linkage dependency with external crates.
861 Rather, [`extern crate` declarations](#extern-crate-declarations) declare linkage dependencies.
863 Use declarations support a number of convenient shortcuts:
865 * Rebinding the target name as a new local name, using the syntax `use x = p::q::r;`.
866 * Simultaneously binding a list of paths differing only in their final element,
867 using the glob-like brace syntax `use a::b::{c,d,e,f};`
868 * Binding all paths matching a given prefix, using the asterisk wildcard syntax `use a::b::*;`
870 An example of `use` declarations:
873 use std::iter::range_step;
874 use std::option::{Some, None};
879 // Equivalent to 'std::iter::range_step(0, 10, 2);'
880 range_step(0, 10, 2);
882 // Equivalent to 'foo(~[std::option::Some(1.0), std::option::None]);'
883 foo(~[Some(1.0), None]);
887 Like items, `use` declarations are private to the containing module, by default.
888 Also like items, a `use` declaration can be public, if qualified by the `pub` keyword.
889 Such a `use` declaration serves to _re-export_ a name.
890 A public `use` declaration can therefore _redirect_ some public name to a different target definition:
891 even a definition with a private canonical path, inside a different module.
892 If a sequence of such redirections form a cycle or cannot be resolved unambiguously,
893 they represent a compile-time error.
895 An example of re-exporting:
900 pub use quux::foo::*;
909 In this example, the module `quux` re-exports all of the public names defined in `foo`.
911 Also note that the paths contained in `use` items are relative to the crate root.
912 So, in the previous example, the `use` refers to `quux::foo::*`, and not simply to `foo::*`.
913 This also means that top-level module declarations should be at the crate root if direct usage
914 of the declared modules within `use` items is desired. It is also possible to use `self` and `super`
915 at the beginning of a `use` item to refer to the current and direct parent modules respectively.
916 All rules regarding accessing declared modules in `use` declarations applies to both module declarations
917 and `extern crate` declarations.
919 An example of what will and will not work for `use` items:
922 # #![allow(unused_imports)]
923 use foo::native::start; // good: foo is at the root of the crate
924 use foo::baz::foobaz; // good: foo is at the root of the crate
929 use foo::native::start; // good: foo is at crate root
930 // use native::start; // bad: native is not at the crate root
931 use self::baz::foobaz; // good: self refers to module 'foo'
932 use foo::bar::foobar; // good: foo is at crate root
939 use super::bar::foobar; // good: super refers to module 'foo'
949 A _function item_ defines a sequence of [statements](#statements) and an optional final [expression](#expressions), along with a name and a set of parameters.
950 Functions are declared with the keyword `fn`.
951 Functions declare a set of *input* [*slots*](#memory-slots) as parameters, through which the caller passes arguments into the function, and an *output* [*slot*](#memory-slots) through which the function passes results back to the caller.
953 A function may also be copied into a first class *value*, in which case the
954 value has the corresponding [*function type*](#function-types), and can be
955 used otherwise exactly as a function item (with a minor additional cost of
956 calling the function indirectly).
958 Every control path in a function logically ends with a `return` expression or a
959 diverging expression. If the outermost block of a function has a
960 value-producing expression in its final-expression position, that expression
961 is interpreted as an implicit `return` expression applied to the
964 An example of a function:
967 fn add(x: int, y: int) -> int {
972 As with `let` bindings, function arguments are irrefutable patterns,
973 so any pattern that is valid in a let binding is also valid as an argument.
976 fn first((value, _): (int, int)) -> int { value }
980 #### Generic functions
982 A _generic function_ allows one or more _parameterized types_ to
983 appear in its signature. Each type parameter must be explicitly
984 declared, in an angle-bracket-enclosed, comma-separated list following
988 fn iter<T>(seq: &[T], f: |T|) {
989 for elt in seq.iter() { f(elt); }
991 fn map<T, U>(seq: &[T], f: |T| -> U) -> ~[U] {
993 for elt in seq.iter() { acc.push(f(elt)); }
998 Inside the function signature and body, the name of the type parameter
999 can be used as a type name.
1001 When a generic function is referenced, its type is instantiated based
1002 on the context of the reference. For example, calling the `iter`
1003 function defined above on `[1, 2]` will instantiate type parameter `T`
1004 with `int`, and require the closure parameter to have type
1007 The type parameters can also be explicitly supplied in a trailing
1008 [path](#paths) component after the function name. This might be necessary
1009 if there is not sufficient context to determine the type parameters. For
1010 example, `mem::size_of::<u32>() == 4`.
1012 Since a parameter type is opaque to the generic function, the set of
1013 operations that can be performed on it is limited. Values of parameter
1014 type can only be moved, not copied.
1017 fn id<T>(x: T) -> T { x }
1020 Similarly, [trait](#traits) bounds can be specified for type
1021 parameters to allow methods with that trait to be called on values
1027 Unsafe operations are those that potentially violate the memory-safety guarantees of Rust's static semantics.
1029 The following language level features cannot be used in the safe subset of Rust:
1031 - Dereferencing a [raw pointer](#pointer-types).
1032 - Calling an unsafe function (including an intrinsic or foreign function).
1034 ##### Unsafe functions
1036 Unsafe functions are functions that are not safe in all contexts and/or for all possible inputs.
1037 Such a function must be prefixed with the keyword `unsafe`.
1041 A block of code can also be prefixed with the `unsafe` keyword, to permit calling `unsafe` functions
1042 or dereferencing raw pointers within a safe function.
1044 When a programmer has sufficient conviction that a sequence of potentially unsafe operations is
1045 actually safe, they can encapsulate that sequence (taken as a whole) within an `unsafe` block. The
1046 compiler will consider uses of such code safe, in the surrounding context.
1048 Unsafe blocks are used to wrap foreign libraries, make direct use of hardware or implement features
1049 not directly present in the language. For example, Rust provides the language features necessary to
1050 implement memory-safe concurrency in the language but the implementation of tasks and message
1051 passing is in the standard library.
1053 Rust's type system is a conservative approximation of the dynamic safety requirements, so in some
1054 cases there is a performance cost to using safe code. For example, a doubly-linked list is not a
1055 tree structure and can only be represented with managed or reference-counted pointers in safe code.
1056 By using `unsafe` blocks to represent the reverse links as raw pointers, it can be implemented with
1057 only owned pointers.
1059 ##### Behavior considered unsafe
1061 This is a list of behavior which is forbidden in all Rust code. Type checking provides the guarantee
1062 that these issues are never caused by safe code. An `unsafe` block or function is responsible for
1063 never invoking this behaviour or exposing an API making it possible for it to occur in safe code.
1066 * Dereferencing a null/dangling raw pointer
1067 * Mutating an immutable value/reference
1068 * Reads of [undef](http://llvm.org/docs/LangRef.html#undefined-values) (uninitialized) memory
1069 * Breaking the [pointer aliasing rules](http://llvm.org/docs/LangRef.html#pointer-aliasing-rules)
1070 with raw pointers (a subset of the rules used by C)
1071 * Invoking undefined behavior via compiler intrinsics:
1072 * Indexing outside of the bounds of an object with `std::ptr::offset` (`offset` intrinsic), with
1073 the exception of one byte past the end which is permitted.
1074 * Using `std::ptr::copy_nonoverlapping_memory` (`memcpy32`/`memcpy64` instrinsics) on
1076 * Invalid values in primitive types, even in private fields/locals:
1077 * Dangling/null pointers in non-raw pointers, or slices
1078 * A value other than `false` (0) or `true` (1) in a `bool`
1079 * A discriminant in an `enum` not included in the type definition
1080 * A value in a `char` which is a surrogate or above `char::MAX`
1081 * non-UTF-8 byte sequences in a `str`
1083 ##### Behaviour not considered unsafe
1085 This is a list of behaviour not considered *unsafe* in Rust terms, but that may be undesired.
1088 * Reading data from private fields (`std::repr`, `format!("{:?}", x)`)
1089 * Leaks due to reference count cycles, even in the global heap
1090 * Exiting without calling destructors
1092 * Accessing/modifying the file system
1093 * Unsigned integer overflow (well-defined as wrapping)
1094 * Signed integer overflow (well-defined as two's complement representation wrapping)
1096 #### Diverging functions
1098 A special kind of function can be declared with a `!` character where the
1099 output slot type would normally be. For example:
1102 fn my_err(s: &str) -> ! {
1108 We call such functions "diverging" because they never return a value to the
1109 caller. Every control path in a diverging function must end with a
1110 `fail!()` or a call to another diverging function on every
1111 control path. The `!` annotation does *not* denote a type. Rather, the result
1112 type of a diverging function is a special type called $\bot$ ("bottom") that
1113 unifies with any type. Rust has no syntax for $\bot$.
1115 It might be necessary to declare a diverging function because as mentioned
1116 previously, the typechecker checks that every control path in a function ends
1117 with a [`return`](#return-expressions) or diverging expression. So, if `my_err`
1118 were declared without the `!` annotation, the following code would not
1122 # fn my_err(s: &str) -> ! { fail!() }
1124 fn f(i: int) -> int {
1129 my_err("Bad number!");
1134 This will not compile without the `!` annotation on `my_err`,
1135 since the `else` branch of the conditional in `f` does not return an `int`,
1136 as required by the signature of `f`.
1137 Adding the `!` annotation to `my_err` informs the typechecker that,
1138 should control ever enter `my_err`, no further type judgments about `f` need to hold,
1139 since control will never resume in any context that relies on those judgments.
1140 Thus the return type on `f` only needs to reflect the `if` branch of the conditional.
1143 #### Extern functions
1145 Extern functions are part of Rust's foreign function interface,
1146 providing the opposite functionality to [external blocks](#external-blocks).
1147 Whereas external blocks allow Rust code to call foreign code,
1148 extern functions with bodies defined in Rust code _can be called by foreign
1149 code_. They are defined in the same way as any other Rust function,
1150 except that they have the `extern` modifier.
1153 // Declares an extern fn, the ABI defaults to "C"
1154 extern fn new_vec() -> ~[int] { ~[] }
1156 // Declares an extern fn with "stdcall" ABI
1157 extern "stdcall" fn new_vec_stdcall() -> ~[int] { ~[] }
1160 Unlike normal functions, extern fns have an `extern "ABI" fn()`.
1161 This is the same type as the functions declared in an extern
1165 # extern fn new_vec() -> ~[int] { ~[] }
1166 let fptr: extern "C" fn() -> ~[int] = new_vec;
1169 Extern functions may be called directly from Rust code as Rust uses large,
1170 contiguous stack segments like C.
1172 ### Type definitions
1174 A _type definition_ defines a new name for an existing [type](#types). Type
1175 definitions are declared with the keyword `type`. Every value has a single,
1176 specific type; the type-specified aspects of a value include:
1178 * Whether the value is composed of sub-values or is indivisible.
1179 * Whether the value represents textual or numerical information.
1180 * Whether the value represents integral or floating-point information.
1181 * The sequence of memory operations required to access the value.
1182 * The [kind](#type-kinds) of the type.
1184 For example, the type `(u8, u8)` defines the set of immutable values that are composite pairs,
1185 each containing two unsigned 8-bit integers accessed by pattern-matching and laid out in memory with the `x` component preceding the `y` component.
1189 A _structure_ is a nominal [structure type](#structure-types) defined with the keyword `struct`.
1191 An example of a `struct` item and its use:
1194 struct Point {x: int, y: int}
1195 let p = Point {x: 10, y: 11};
1199 A _tuple structure_ is a nominal [tuple type](#tuple-types), also defined with the keyword `struct`.
1203 struct Point(int, int);
1204 let p = Point(10, 11);
1205 let px: int = match p { Point(x, _) => x };
1208 A _unit-like struct_ is a structure without any fields, defined by leaving off the list of fields entirely.
1209 Such types will have a single value, just like the [unit value `()`](#unit-and-boolean-literals) of the unit type.
1214 let c = [Cookie, Cookie, Cookie, Cookie];
1219 An _enumeration_ is a simultaneous definition of a nominal [enumerated type](#enumerated-types) as well as a set of *constructors*,
1220 that can be used to create or pattern-match values of the corresponding enumerated type.
1222 Enumerations are declared with the keyword `enum`.
1224 An example of an `enum` item and its use:
1232 let mut a: Animal = Dog;
1236 Enumeration constructors can have either named or unnamed fields:
1241 Cat { name: ~str, weight: f64 }
1244 let mut a: Animal = Dog(~"Cocoa", 37.2);
1245 a = Cat{ name: ~"Spotty", weight: 2.7 };
1248 In this example, `Cat` is a _struct-like enum variant_,
1249 whereas `Dog` is simply called an enum variant.
1253 ~~~~ {.notrust .ebnf .gram}
1254 static_item : "static" ident ':' type '=' expr ';' ;
1257 A *static item* is a named _constant value_ stored in the global data section of a crate.
1258 Immutable static items are stored in the read-only data section.
1259 The constant value bound to a static item is, like all constant values, evaluated at compile time.
1260 Static items have the `static` lifetime, which outlives all other lifetimes in a Rust program.
1261 Static items are declared with the `static` keyword.
1262 A static item must have a _constant expression_ giving its definition.
1264 Static items must be explicitly typed.
1265 The type may be ```bool```, ```char```, a number, or a type derived from those primitive types.
1266 The derived types are references with the `static` lifetime,
1267 fixed-size arrays, tuples, and structs.
1270 static BIT1: uint = 1 << 0;
1271 static BIT2: uint = 1 << 1;
1273 static BITS: [uint, ..2] = [BIT1, BIT2];
1274 static STRING: &'static str = "bitstring";
1276 struct BitsNStrings<'a> {
1277 mybits: [uint, ..2],
1281 static bits_n_strings: BitsNStrings<'static> = BitsNStrings {
1287 #### Mutable statics
1289 If a static item is declared with the ```mut``` keyword, then it is allowed to
1290 be modified by the program. One of Rust's goals is to make concurrency bugs hard
1291 to run into, and this is obviously a very large source of race conditions or
1292 other bugs. For this reason, an ```unsafe``` block is required when either
1293 reading or writing a mutable static variable. Care should be taken to ensure
1294 that modifications to a mutable static are safe with respect to other tasks
1295 running in the same process.
1297 Mutable statics are still very useful, however. They can be used with C
1298 libraries and can also be bound from C libraries (in an ```extern``` block).
1301 # fn atomic_add(_: &mut uint, _: uint) -> uint { 2 }
1303 static mut LEVELS: uint = 0;
1305 // This violates the idea of no shared state, and this doesn't internally
1306 // protect against races, so this function is `unsafe`
1307 unsafe fn bump_levels_unsafe1() -> uint {
1313 // Assuming that we have an atomic_add function which returns the old value,
1314 // this function is "safe" but the meaning of the return value may not be what
1315 // callers expect, so it's still marked as `unsafe`
1316 unsafe fn bump_levels_unsafe2() -> uint {
1317 return atomic_add(&mut LEVELS, 1);
1323 A _trait_ describes a set of method types.
1325 Traits can include default implementations of methods,
1326 written in terms of some unknown [`self` type](#self-types);
1327 the `self` type may either be completely unspecified,
1328 or constrained by some other trait.
1330 Traits are implemented for specific types through separate [implementations](#implementations).
1333 # type Surface = int;
1334 # type BoundingBox = int;
1337 fn draw(&self, Surface);
1338 fn bounding_box(&self) -> BoundingBox;
1342 This defines a trait with two methods.
1343 All values that have [implementations](#implementations) of this trait in scope can have their `draw` and `bounding_box` methods called,
1344 using `value.bounding_box()` [syntax](#method-call-expressions).
1346 Type parameters can be specified for a trait to make it generic.
1347 These appear after the trait name, using the same syntax used in [generic functions](#generic-functions).
1351 fn len(&self) -> uint;
1352 fn elt_at(&self, n: uint) -> T;
1353 fn iter(&self, |T|);
1357 Generic functions may use traits as _bounds_ on their type parameters.
1358 This will have two effects: only types that have the trait may instantiate the parameter,
1359 and within the generic function,
1360 the methods of the trait can be called on values that have the parameter's type.
1364 # type Surface = int;
1365 # trait Shape { fn draw(&self, Surface); }
1367 fn draw_twice<T: Shape>(surface: Surface, sh: T) {
1373 Traits also define an [object type](#object-types) with the same name as the trait.
1374 Values of this type are created by [casting](#type-cast-expressions) pointer values
1375 (pointing to a type for which an implementation of the given trait is in scope)
1376 to pointers to the trait name, used as a type.
1380 # impl Shape for int { }
1383 let myshape: ~Shape = ~mycircle as ~Shape;
1386 The resulting value is a managed box containing the value that was cast,
1387 along with information that identifies the methods of the implementation that was used.
1388 Values with a trait type can have [methods called](#method-call-expressions) on them,
1389 for any method in the trait,
1390 and can be used to instantiate type parameters that are bounded by the trait.
1392 Trait methods may be static,
1393 which means that they lack a `self` argument.
1394 This means that they can only be called with function call syntax (`f(x)`)
1395 and not method call syntax (`obj.f()`).
1396 The way to refer to the name of a static method is to qualify it with the trait name,
1397 treating the trait name like a module.
1402 fn from_int(n: int) -> Self;
1405 fn from_int(n: int) -> f64 { n as f64 }
1407 let x: f64 = Num::from_int(42);
1410 Traits may inherit from other traits. For example, in
1413 trait Shape { fn area() -> f64; }
1414 trait Circle : Shape { fn radius() -> f64; }
1417 the syntax `Circle : Shape` means that types that implement `Circle` must also have an implementation for `Shape`.
1418 Multiple supertraits are separated by spaces, `trait Circle : Shape Eq { }`.
1419 In an implementation of `Circle` for a given type `T`, methods can refer to `Shape` methods,
1420 since the typechecker checks that any type with an implementation of `Circle` also has an implementation of `Shape`.
1422 In type-parameterized functions,
1423 methods of the supertrait may be called on values of subtrait-bound type parameters.
1424 Referring to the previous example of `trait Circle : Shape`:
1427 # trait Shape { fn area(&self) -> f64; }
1428 # trait Circle : Shape { fn radius(&self) -> f64; }
1429 fn radius_times_area<T: Circle>(c: T) -> f64 {
1430 // `c` is both a Circle and a Shape
1431 c.radius() * c.area()
1435 Likewise, supertrait methods may also be called on trait objects.
1438 # trait Shape { fn area(&self) -> f64; }
1439 # trait Circle : Shape { fn radius(&self) -> f64; }
1440 # impl Shape for int { fn area(&self) -> f64 { 0.0 } }
1441 # impl Circle for int { fn radius(&self) -> f64 { 0.0 } }
1444 let mycircle: Circle = ~mycircle as ~Circle;
1445 let nonsense = mycircle.radius() * mycircle.area();
1450 An _implementation_ is an item that implements a [trait](#traits) for a specific type.
1452 Implementations are defined with the keyword `impl`.
1455 # struct Point {x: f64, y: f64};
1456 # type Surface = int;
1457 # struct BoundingBox {x: f64, y: f64, width: f64, height: f64};
1458 # trait Shape { fn draw(&self, Surface); fn bounding_box(&self) -> BoundingBox; }
1459 # fn do_draw_circle(s: Surface, c: Circle) { }
1466 impl Shape for Circle {
1467 fn draw(&self, s: Surface) { do_draw_circle(s, *self); }
1468 fn bounding_box(&self) -> BoundingBox {
1469 let r = self.radius;
1470 BoundingBox{x: self.center.x - r, y: self.center.y - r,
1471 width: 2.0 * r, height: 2.0 * r}
1476 It is possible to define an implementation without referring to a trait.
1477 The methods in such an implementation can only be used
1478 as direct calls on the values of the type that the implementation targets.
1479 In such an implementation, the trait type and `for` after `impl` are omitted.
1480 Such implementations are limited to nominal types (enums, structs),
1481 and the implementation must appear in the same module or a sub-module as the `self` type.
1483 When a trait _is_ specified in an `impl`,
1484 all methods declared as part of the trait must be implemented,
1485 with matching types and type parameter counts.
1487 An implementation can take type parameters,
1488 which can be different from the type parameters taken by the trait it implements.
1489 Implementation parameters are written after the `impl` keyword.
1494 impl<T> Seq<T> for ~[T] {
1497 impl Seq<bool> for u32 {
1498 /* Treat the integer as a sequence of bits */
1504 ~~~~ {.notrust .ebnf .gram}
1505 extern_block_item : "extern" '{' extern_block '}' ;
1506 extern_block : [ foreign_fn ] * ;
1509 External blocks form the basis for Rust's foreign function interface.
1510 Declarations in an external block describe symbols
1511 in external, non-Rust libraries.
1513 Functions within external blocks
1514 are declared in the same way as other Rust functions,
1515 with the exception that they may not have a body
1516 and are instead terminated by a semicolon.
1520 use libc::{c_char, FILE};
1523 fn fopen(filename: *c_char, mode: *c_char) -> *FILE;
1528 Functions within external blocks may be called by Rust code,
1529 just like functions defined in Rust.
1530 The Rust compiler automatically translates
1531 between the Rust ABI and the foreign ABI.
1533 A number of [attributes](#attributes) control the behavior of external
1536 By default external blocks assume that the library they are calling
1537 uses the standard C "cdecl" ABI. Other ABIs may be specified using
1538 an `abi` string, as shown here:
1541 // Interface to the Windows API
1542 extern "stdcall" { }
1545 The `link` attribute allows the name of the library to be specified. When
1546 specified the compiler will attempt to link against the native library of the
1550 #[link(name = "crypto")]
1554 The type of a function declared in an extern block is `extern "abi" fn(A1,
1555 ..., An) -> R`, where `A1...An` are the declared types of its arguments and
1556 `R` is the declared return type.
1558 ## Visibility and Privacy
1560 These two terms are often used interchangeably, and what they are attempting to
1561 convey is the answer to the question "Can this item be used at this location?"
1563 Rust's name resolution operates on a global hierarchy of namespaces. Each level
1564 in the hierarchy can be thought of as some item. The items are one of those
1565 mentioned above, but also include external crates. Declaring or defining a new
1566 module can be thought of as inserting a new tree into the hierarchy at the
1567 location of the definition.
1569 To control whether interfaces can be used across modules, Rust checks each use
1570 of an item to see whether it should be allowed or not. This is where privacy
1571 warnings are generated, or otherwise "you used a private item of another module
1572 and weren't allowed to."
1574 By default, everything in rust is *private*, with one exception. Enum variants
1575 in a `pub` enum are also public by default. You are allowed to alter this
1576 default visibility with the `priv` keyword. When an item is declared as `pub`,
1577 it can be thought of as being accessible to the outside world. For example:
1581 // Declare a private struct
1584 // Declare a public struct with a private field
1589 // Declare a public enum with two public variants
1591 PubliclyAccessibleState,
1592 PubliclyAccessibleState2,
1596 With the notion of an item being either public or private, Rust allows item
1597 accesses in two cases:
1599 1. If an item is public, then it can be used externally through any of its
1601 2. If an item is private, it may be accessed by the current module and its
1604 These two cases are surprisingly powerful for creating module hierarchies
1605 exposing public APIs while hiding internal implementation details. To help
1606 explain, here's a few use cases and what they would entail.
1608 * A library developer needs to expose functionality to crates which link against
1609 their library. As a consequence of the first case, this means that anything
1610 which is usable externally must be `pub` from the root down to the destination
1611 item. Any private item in the chain will disallow external accesses.
1613 * A crate needs a global available "helper module" to itself, but it doesn't
1614 want to expose the helper module as a public API. To accomplish this, the root
1615 of the crate's hierarchy would have a private module which then internally has
1616 a "public api". Because the entire crate is a descendant of the root, then the
1617 entire local crate can access this private module through the second case.
1619 * When writing unit tests for a module, it's often a common idiom to have an
1620 immediate child of the module to-be-tested named `mod test`. This module could
1621 access any items of the parent module through the second case, meaning that
1622 internal implementation details could also be seamlessly tested from the child
1625 In the second case, it mentions that a private item "can be accessed" by the
1626 current module and its descendants, but the exact meaning of accessing an item
1627 depends on what the item is. Accessing a module, for example, would mean looking
1628 inside of it (to import more items). On the other hand, accessing a function
1629 would mean that it is invoked. Additionally, path expressions and import
1630 statements are considered to access an item in the sense that the
1631 import/expression is only valid if the destination is in the current visibility
1634 Here's an example of a program which exemplifies the three cases outlined above.
1637 // This module is private, meaning that no external crate can access this
1638 // module. Because it is private at the root of this current crate, however, any
1639 // module in the crate may access any publicly visible item in this module.
1640 mod crate_helper_module {
1642 // This function can be used by anything in the current crate
1643 pub fn crate_helper() {}
1645 // This function *cannot* be used by anything else in the crate. It is not
1646 // publicly visible outside of the `crate_helper_module`, so only this
1647 // current module and its descendants may access it.
1648 fn implementation_detail() {}
1651 // This function is "public to the root" meaning that it's available to external
1652 // crates linking against this one.
1653 pub fn public_api() {}
1655 // Similarly to 'public_api', this module is public so external crates may look
1658 use crate_helper_module;
1660 pub fn my_method() {
1661 // Any item in the local crate may invoke the helper module's public
1662 // interface through a combination of the two rules above.
1663 crate_helper_module::crate_helper();
1666 // This function is hidden to any module which is not a descendant of
1668 fn my_implementation() {}
1674 fn test_my_implementation() {
1675 // Because this module is a descendant of `submodule`, it's allowed
1676 // to access private items inside of `submodule` without a privacy
1678 super::my_implementation();
1686 For a rust program to pass the privacy checking pass, all paths must be valid
1687 accesses given the two rules above. This includes all use statements,
1688 expressions, types, etc.
1690 ### Re-exporting and Visibility
1692 Rust allows publicly re-exporting items through a `pub use` directive. Because
1693 this is a public directive, this allows the item to be used in the current
1694 module through the rules above. It essentially allows public access into the
1695 re-exported item. For example, this program is valid:
1698 pub use api = self::implementation;
1700 mod implementation {
1707 This means that any external crate referencing `implementation::f` would receive
1708 a privacy violation, while the path `api::f` would be allowed.
1710 When re-exporting a private item, it can be thought of as allowing the "privacy
1711 chain" being short-circuited through the reexport instead of passing through the
1712 namespace hierarchy as it normally would.
1714 ### Glob imports and Visibility
1716 Currently glob imports are considered an "experimental" language feature. For
1717 sanity purpose along with helping the implementation, glob imports will only
1718 import public items from their destination, not private items.
1720 > **Note:** This is subject to change, glob exports may be removed entirely or
1721 > they could possibly import private items for a privacy error to later be
1722 > issued if the item is used.
1726 ~~~~ {.notrust .ebnf .gram}
1727 attribute : '#' '!' ? '[' attr_list ']' ;
1728 attr_list : attr [ ',' attr_list ]* ;
1729 attr : ident [ '=' literal
1730 | '(' attr_list ')' ] ? ;
1733 Static entities in Rust -- crates, modules and items -- may have _attributes_
1734 applied to them. Attributes in Rust are modeled on Attributes in ECMA-335,
1735 with the syntax coming from ECMA-334 (C#). An attribute is a general,
1736 free-form metadatum that is interpreted according to name, convention, and
1737 language and compiler version. Attributes may appear as any of:
1739 * A single identifier, the attribute name
1740 * An identifier followed by the equals sign '=' and a literal, providing a
1742 * An identifier followed by a parenthesized list of sub-attribute arguments
1744 Attributes with a bang ("!") after the hash ("#") apply to the item that the
1745 attribute is declared within. Attributes that do not have a bang after the
1746 hash apply to the item that follows the attribute.
1748 An example of attributes:
1751 // General metadata applied to the enclosing module or crate.
1754 // A function marked as a unit test
1760 // A conditionally-compiled module
1761 #[cfg(target_os="linux")]
1766 // A lint attribute used to suppress a warning/error
1767 #[allow(non_camel_case_types)]
1771 > **Note:** At some point in the future, the compiler will distinguish between
1772 > language-reserved and user-available attributes. Until then, there is
1773 > effectively no difference between an attribute handled by a loadable syntax
1774 > extension and the compiler.
1776 ### Crate-only attributes
1778 - `crate_id` - specify the this crate's crate ID.
1779 - `crate_type` - see [linkage](#linkage).
1780 - `feature` - see [compiler features](#compiler-features).
1781 - `no_main` - disable emitting the `main` symbol. Useful when some other
1782 object being linked to defines `main`.
1783 - `no_start` - disable linking to the `native` crate, which specifies the
1784 "start" language item.
1785 - `no_std` - disable linking to the `std` crate.
1787 ### Module-only attributes
1789 - `macro_escape` - macros defined in this module will be visible in the
1790 module's parent, after this module has been included.
1791 - `no_implicit_prelude` - disable injecting `use std::prelude::*` in this
1793 - `path` - specifies the file to load the module from. `#[path="foo.rs"] mod
1794 bar;` is equivalent to `mod bar { /* contents of foo.rs */ }`. The path is
1795 taken relative to the directory that the current module is in.
1797 ### Function-only attributes
1799 - `macro_registrar` - when using loadable syntax extensions, mark this
1800 function as the registration point for the current crate's syntax
1802 - `main` - indicates that this function should be passed to the entry point,
1803 rather than the function in the crate root named `main`.
1804 - `start` - indicates that this function should be used as the entry point,
1805 overriding the "start" language item. See the "start" [language
1806 item](#language-items) for more details.
1808 ### Static-only attributes
1810 - `address_insignificant` - references to this static may alias with
1811 references to other statics, potentially of unrelated type.
1812 - `thread_local` - on a `static mut`, this signals that the value of this
1813 static may change depending on the current thread. The exact consequences of
1814 this are implementation-defined.
1818 On an `extern` block, the following attributes are interpreted:
1820 - `link_args` - specify arguments to the linker, rather than just the library
1821 name and type. This is feature gated and the exact behavior is
1822 implementation-defined (due to variety of linker invocation syntax).
1823 - `link` - indicate that a native library should be linked to for the
1824 declarations in this block to be linked correctly. See [external
1825 blocks](#external-blocks)
1827 On declarations inside an `extern` block, the following attributes are
1830 - `link_name` - the name of the symbol that this function or static should be
1832 - `linkage` - on a static, this specifies the [linkage
1833 type](http://llvm.org/docs/LangRef.html#linkage-types).
1835 ### Miscellaneous attributes
1837 - `link_section` - on statics and functions, this specifies the section of the
1838 object file that this item's contents will be placed into.
1839 - `macro_export` - export a macro for cross-crate usage.
1840 - `no_mangle` - on any item, do not apply the standard name mangling. Set the
1841 symbol for this item to its identifier.
1842 - `packed` - on structs or enums, eliminate any padding that would be used to
1844 - `repr` - on C-like enums, this sets the underlying type used for
1845 representation. Useful for FFI. Takes one argument, which is the primitive
1846 type this enum should be represented for, or `C`, which specifies that it
1847 should be the default `enum` size of the C ABI for that platform. Note that
1848 enum representation in C is undefined, and this may be incorrect when the C
1849 code is compiled with certain flags.
1850 - `simd` - on certain tuple structs, derive the arithmetic operators, which
1851 lower to the target's SIMD instructions, if any.
1852 - `static_assert` - on statics whose type is `bool`, terminates compilation
1853 with an error if it is not initialized to `true`.
1854 - `unsafe_destructor` - allow implementations of the "drop" language item
1855 where the type it is implemented for does not implement the "send" language
1857 - `unsafe_no_drop_flag` - on structs, remove the flag that prevents
1858 destructors from being run twice. Destructors might be run multiple times on
1859 the same object with this attribute.
1861 ### Conditional compilation
1863 Sometimes one wants to have different compiler outputs from the same code,
1864 depending on build target, such as targeted operating system, or to enable
1867 There are two kinds of configuration options, one that is either defined or not
1868 (`#[cfg(foo)]`), and the other that contains a string that can be checked
1869 against (`#[cfg(bar = "baz")]` (currently only compiler-defined configuration
1870 options can have the latter form).
1873 // The function is only included in the build when compiling for OSX
1874 #[cfg(target_os = "macos")]
1879 // This function is only included when either foo or bar is defined
1882 fn needs_foo_or_bar() {
1886 // This function is only included when compiling for a unixish OS with a 32-bit
1888 #[cfg(unix, target_word_size = "32")]
1889 fn on_32bit_unix() {
1894 This illustrates some conditional compilation can be achieved using the
1895 `#[cfg(...)]` attribute. Note that `#[cfg(foo, bar)]` is a condition that needs
1896 both `foo` and `bar` to be defined while `#[cfg(foo)] #[cfg(bar)]` only needs
1897 one of `foo` and `bar` to be defined (this resembles in the disjunctive normal
1898 form). Additionally, one can reverse a condition by enclosing it in a
1899 `not(...)`, like e. g. `#[cfg(not(target_os = "win32"))]`.
1901 The following configurations must be defined by the implementation:
1903 * `target_arch = "..."`. Target CPU architecture, such as `"x86"`, `"x86_64"`
1904 `"mips"`, or `"arm"`.
1905 * `target_endian = "..."`. Endianness of the target CPU, either `"little"` or
1907 * `target_family = "..."`. Operating system family of the target, e. g.
1908 `"unix"` or `"windows"`. The value of this configuration option is defined as
1909 a configuration itself, like `unix` or `windows`.
1910 * `target_os = "..."`. Operating system of the target, examples include
1911 `"win32"`, `"macos"`, `"linux"`, `"android"` or `"freebsd"`.
1912 * `target_word_size = "..."`. Target word size in bits. This is set to `"32"`
1913 for targets with 32-bit pointers, and likewise set to `"64"` for 64-bit
1915 * `unix`. See `target_family`.
1916 * `windows`. See `target_family`.
1918 ### Lint check attributes
1920 A lint check names a potentially undesirable coding pattern, such as
1921 unreachable code or omitted documentation, for the static entity to
1922 which the attribute applies.
1924 For any lint check `C`:
1926 * `warn(C)` warns about violations of `C` but continues compilation,
1927 * `deny(C)` signals an error after encountering a violation of `C`,
1928 * `allow(C)` overrides the check for `C` so that violations will go
1930 * `forbid(C)` is the same as `deny(C)`, but also forbids changing the lint
1933 The lint checks supported by the compiler can be found via `rustc -W help`,
1934 along with their default settings.
1938 // Missing documentation is ignored here
1939 #[allow(missing_doc)]
1940 pub fn undocumented_one() -> int { 1 }
1942 // Missing documentation signals a warning here
1943 #[warn(missing_doc)]
1944 pub fn undocumented_too() -> int { 2 }
1946 // Missing documentation signals an error here
1947 #[deny(missing_doc)]
1948 pub fn undocumented_end() -> int { 3 }
1952 This example shows how one can use `allow` and `warn` to toggle
1953 a particular check on and off.
1956 #[warn(missing_doc)]
1958 #[allow(missing_doc)]
1960 // Missing documentation is ignored here
1961 pub fn undocumented_one() -> int { 1 }
1963 // Missing documentation signals a warning here,
1964 // despite the allow above.
1965 #[warn(missing_doc)]
1966 pub fn undocumented_two() -> int { 2 }
1969 // Missing documentation signals a warning here
1970 pub fn undocumented_too() -> int { 3 }
1974 This example shows how one can use `forbid` to disallow uses
1975 of `allow` for that lint check.
1978 #[forbid(missing_doc)]
1980 // Attempting to toggle warning signals an error here
1981 #[allow(missing_doc)]
1983 pub fn undocumented_too() -> int { 2 }
1989 Some primitive Rust operations are defined in Rust code, rather than being
1990 implemented directly in C or assembly language. The definitions of these
1991 operations have to be easy for the compiler to find. The `lang` attribute
1992 makes it possible to declare these operations. For example, the `str` module
1993 in the Rust standard library defines the string equality function:
1997 pub fn eq_slice(a: &str, b: &str) -> bool {
2002 The name `str_eq` has a special meaning to the Rust compiler,
2003 and the presence of this definition means that it will use this definition
2004 when generating calls to the string equality function.
2006 A complete list of the built-in language items follows:
2008 #### Built-in Traits
2011 : Able to be sent across task boundaries.
2013 : Has a size known at compile time.
2015 : Types that do not move ownership when used by-value.
2017 : Able to be safely shared between tasks when aliased.
2023 These language items are traits:
2026 : Elements can be added (for example, integers and floats).
2028 : Elements can be subtracted.
2030 : Elements can be multiplied.
2032 : Elements have a division operation.
2034 : Elements have a remainder operation.
2036 : Elements can be negated arithmetically.
2038 : Elements can be negated logically.
2040 : Elements have an exclusive-or operation.
2042 : Elements have a bitwise `and` operation.
2044 : Elements have a bitwise `or` operation.
2046 : Elements have a left shift operation.
2048 : Elements have a right shift operation.
2050 : Elements can be indexed.
2052 : Elements can be compared for equality.
2054 : Elements have a partial ordering.
2056 : `*` can be applied, yielding a reference to another type
2058 : `*` can be applied, yielding a mutable reference to another type
2061 These are functions:
2064 : Compare two strings (`&str`) for equality.
2066 : Compare two owned strings (`~str`) for equality.
2068 : Return a new unique string
2069 containing a copy of the contents of a unique string.
2074 : A type whose contents can be mutated through an immutable reference
2076 : The type returned by the `type_id` intrinsic.
2080 These types help drive the compiler's analysis
2083 : The type parameter should be considered covariant
2084 * `contravariant_type`
2085 : The type parameter should be considered contravariant
2087 : The type parameter should be considered invariant
2088 * `covariant_lifetime`
2089 : The lifetime parameter should be considered covariant
2090 * `contravariant_lifetime`
2091 : The lifetime parameter should be considered contravariant
2092 * `invariant_lifetime`
2093 : The lifetime parameter should be considered invariant
2095 : This type does not implement "send", even if eligible
2097 : This type does not implement "copy", even if eligible
2099 : This type does not implement "share", even if eligible
2101 : This type implements "managed"
2104 : Abort the program with an error.
2105 * `fail_bounds_check`
2106 : Abort the program with a bounds check error.
2108 : Allocate memory on the exchange heap.
2110 : Free memory that was allocated on the exchange heap.
2112 : Allocate memory on the managed heap.
2114 : Free memory that was allocated on the managed heap.
2116 > **Note:** This list is likely to become out of date. We should auto-generate it
2117 > from `librustc/middle/lang_items.rs`.
2119 ### Inline attributes
2121 The inline attribute is used to suggest to the compiler to perform an inline
2122 expansion and place a copy of the function in the caller rather than generating
2123 code to call the function where it is defined.
2125 The compiler automatically inlines functions based on internal heuristics.
2126 Incorrectly inlining functions can actually making the program slower, so it
2127 should be used with care.
2129 `#[inline]` and `#[inline(always)]` always causes the function to be serialized
2130 into crate metadata to allow cross-crate inlining.
2132 There are three different types of inline attributes:
2134 * `#[inline]` hints the compiler to perform an inline expansion.
2135 * `#[inline(always)]` asks the compiler to always perform an inline expansion.
2136 * `#[inline(never)]` asks the compiler to never perform an inline expansion.
2140 The `deriving` attribute allows certain traits to be automatically
2141 implemented for data structures. For example, the following will
2142 create an `impl` for the `Eq` and `Clone` traits for `Foo`, the type
2143 parameter `T` will be given the `Eq` or `Clone` constraints for the
2147 #[deriving(Eq, Clone)]
2154 The generated `impl` for `Eq` is equivalent to
2157 # struct Foo<T> { a: int, b: T }
2158 impl<T: Eq> Eq for Foo<T> {
2159 fn eq(&self, other: &Foo<T>) -> bool {
2160 self.a == other.a && self.b == other.b
2163 fn ne(&self, other: &Foo<T>) -> bool {
2164 self.a != other.a || self.b != other.b
2169 Supported traits for `deriving` are:
2171 * Comparison traits: `Eq`, `TotalEq`, `Ord`, `TotalOrd`.
2172 * Serialization: `Encodable`, `Decodable`. These require `serialize`.
2173 * `Clone`, to create `T` from `&T` via a copy.
2174 * `Hash`, to iterate over the bytes in a data type.
2175 * `Rand`, to create a random instance of a data type.
2176 * `Default`, to create an empty instance of a data type.
2177 * `Zero`, to create an zero instance of a numeric data type.
2178 * `FromPrimitive`, to create an instance from a numeric primitive.
2179 * `Show`, to format a value using the `{}` formatter.
2183 One can indicate the stability of an API using the following attributes:
2185 * `deprecated`: This item should no longer be used, e.g. it has been
2186 replaced. No guarantee of backwards-compatibility.
2187 * `experimental`: This item was only recently introduced or is
2188 otherwise in a state of flux. It may change significantly, or even
2189 be removed. No guarantee of backwards-compatibility.
2190 * `unstable`: This item is still under development, but requires more
2191 testing to be considered stable. No guarantee of backwards-compatibility.
2192 * `stable`: This item is considered stable, and will not change
2193 significantly. Guarantee of backwards-compatibility.
2194 * `frozen`: This item is very stable, and is unlikely to
2195 change. Guarantee of backwards-compatibility.
2196 * `locked`: This item will never change unless a serious bug is
2197 found. Guarantee of backwards-compatibility.
2199 These levels are directly inspired by
2200 [Node.js' "stability index"](http://nodejs.org/api/documentation.html).
2202 There are lints for disallowing items marked with certain levels:
2203 `deprecated`, `experimental` and `unstable`; the first two will warn
2204 by default. Items with not marked with a stability are considered to
2205 be unstable for the purposes of the lint. One can give an optional
2206 string that will be displayed when the lint flags the use of an item.
2211 #[deprecated="replaced by `best`"]
2213 // delete everything
2217 // delete fewer things
2226 bad(); // "warning: use of deprecated item: replaced by `best`"
2228 better(); // "warning: use of unmarked item"
2230 best(); // no warning
2234 > **Note:** Currently these are only checked when applied to
2235 > individual functions, structs, methods and enum variants, *not* to
2236 > entire modules, traits, impls or enums themselves.
2238 ### Compiler Features
2240 Certain aspects of Rust may be implemented in the compiler, but they're not
2241 necessarily ready for every-day use. These features are often of "prototype
2242 quality" or "almost production ready", but may not be stable enough to be
2243 considered a full-fleged language feature.
2245 For this reason, Rust recognizes a special crate-level attribute of the form:
2248 #![feature(feature1, feature2, feature3)]
2251 This directive informs the compiler that the feature list: `feature1`,
2252 `feature2`, and `feature3` should all be enabled. This is only recognized at a
2253 crate-level, not at a module-level. Without this directive, all features are
2254 considered off, and using the features will result in a compiler error.
2256 The currently implemented features of the reference compiler are:
2258 * `macro_rules` - The definition of new macros. This does not encompass
2259 macro-invocation, that is always enabled by default, this only
2260 covers the definition of new macros. There are currently
2261 various problems with invoking macros, how they interact with
2262 their environment, and possibly how they are used outside of
2263 location in which they are defined. Macro definitions are
2264 likely to change slightly in the future, so they are currently
2265 hidden behind this feature.
2267 * `globs` - Importing everything in a module through `*`. This is currently a
2268 large source of bugs in name resolution for Rust, and it's not clear
2269 whether this will continue as a feature or not. For these reasons,
2270 the glob import statement has been hidden behind this feature flag.
2272 * `struct_variant` - Structural enum variants (those with named fields). It is
2273 currently unknown whether this style of enum variant is as
2274 fully supported as the tuple-forms, and it's not certain
2275 that this style of variant should remain in the language.
2276 For now this style of variant is hidden behind a feature
2279 * `once_fns` - Onceness guarantees a closure is only executed once. Defining a
2280 closure as `once` is unlikely to be supported going forward. So
2281 they are hidden behind this feature until they are to be removed.
2283 * `managed_boxes` - Usage of `@` pointers is gated due to many
2284 planned changes to this feature. In the past, this has meant
2285 "a GC pointer", but the current implementation uses
2286 reference counting and will likely change drastically over
2287 time. Additionally, the `@` syntax will no longer be used to
2290 * `asm` - The `asm!` macro provides a means for inline assembly. This is often
2291 useful, but the exact syntax for this feature along with its semantics
2292 are likely to change, so this macro usage must be opted into.
2294 * `non_ascii_idents` - The compiler supports the use of non-ascii identifiers,
2295 but the implementation is a little rough around the
2296 edges, so this can be seen as an experimental feature for
2297 now until the specification of identifiers is fully
2300 * `thread_local` - The usage of the `#[thread_local]` attribute is experimental
2301 and should be seen as unstable. This attribute is used to
2302 declare a `static` as being unique per-thread leveraging
2303 LLVM's implementation which works in concert with the kernel
2304 loader and dynamic linker. This is not necessarily available
2305 on all platforms, and usage of it is discouraged (rust
2306 focuses more on task-local data instead of thread-local
2309 * `link_args` - This attribute is used to specify custom flags to the linker,
2310 but usage is strongly discouraged. The compiler's usage of the
2311 system linker is not guaranteed to continue in the future, and
2312 if the system linker is not used then specifying custom flags
2313 doesn't have much meaning.
2315 If a feature is promoted to a language feature, then all existing programs will
2316 start to receive compilation warnings about #[feature] directives which enabled
2317 the new feature (because the directive is no longer necessary). However, if
2318 a feature is decided to be removed from the language, errors will be issued (if
2319 there isn't a parser error first). The directive in this case is no longer
2320 necessary, and it's likely that existing code will break if the feature isn't
2323 If a unknown feature is found in a directive, it results in a compiler error. An
2324 unknown feature is one which has never been recognized by the compiler.
2326 # Statements and expressions
2328 Rust is _primarily_ an expression language. This means that most forms of
2329 value-producing or effect-causing evaluation are directed by the uniform
2330 syntax category of _expressions_. Each kind of expression can typically _nest_
2331 within each other kind of expression, and rules for evaluation of expressions
2332 involve specifying both the value produced by the expression and the order in
2333 which its sub-expressions are themselves evaluated.
2335 In contrast, statements in Rust serve _mostly_ to contain and explicitly
2336 sequence expression evaluation.
2340 A _statement_ is a component of a block, which is in turn a component of an
2341 outer [expression](#expressions) or [function](#functions).
2343 Rust has two kinds of statement:
2344 [declaration statements](#declaration-statements) and
2345 [expression statements](#expression-statements).
2347 ### Declaration statements
2349 A _declaration statement_ is one that introduces one or more *names* into the enclosing statement block.
2350 The declared names may denote new slots or new items.
2352 #### Item declarations
2354 An _item declaration statement_ has a syntactic form identical to an
2355 [item](#items) declaration within a module. Declaring an item -- a function,
2356 enumeration, structure, type, static, trait, implementation or module -- locally
2357 within a statement block is simply a way of restricting its scope to a narrow
2358 region containing all of its uses; it is otherwise identical in meaning to
2359 declaring the item outside the statement block.
2361 Note: there is no implicit capture of the function's dynamic environment when
2362 declaring a function-local item.
2364 #### Slot declarations
2366 ~~~~ {.notrust .ebnf .gram}
2367 let_decl : "let" pat [':' type ] ? [ init ] ? ';' ;
2368 init : [ '=' ] expr ;
2371 A _slot declaration_ introduces a new set of slots, given by a pattern.
2372 The pattern may be followed by a type annotation, and/or an initializer expression.
2373 When no type annotation is given, the compiler will infer the type,
2374 or signal an error if insufficient type information is available for definite inference.
2375 Any slots introduced by a slot declaration are visible from the point of declaration until the end of the enclosing block scope.
2377 ### Expression statements
2379 An _expression statement_ is one that evaluates an [expression](#expressions)
2380 and ignores its result.
2381 The type of an expression statement `e;` is always `()`, regardless of the type of `e`.
2382 As a rule, an expression statement's purpose is to trigger the effects of evaluating its expression.
2386 An expression may have two roles: it always produces a *value*, and it may have *effects*
2387 (otherwise known as "side effects").
2388 An expression *evaluates to* a value, and has effects during *evaluation*.
2389 Many expressions contain sub-expressions (operands).
2390 The meaning of each kind of expression dictates several things:
2391 * Whether or not to evaluate the sub-expressions when evaluating the expression
2392 * The order in which to evaluate the sub-expressions
2393 * How to combine the sub-expressions' values to obtain the value of the expression.
2395 In this way, the structure of expressions dictates the structure of execution.
2396 Blocks are just another kind of expression,
2397 so blocks, statements, expressions, and blocks again can recursively nest inside each other
2398 to an arbitrary depth.
2400 #### Lvalues, rvalues and temporaries
2402 Expressions are divided into two main categories: _lvalues_ and _rvalues_.
2403 Likewise within each expression, sub-expressions may occur in _lvalue context_ or _rvalue context_.
2404 The evaluation of an expression depends both on its own category and the context it occurs within.
2406 An lvalue is an expression that represents a memory location. These
2407 expressions are [paths](#path-expressions) (which refer to local
2408 variables, function and method arguments, or static variables),
2409 dereferences (`*expr`), [indexing expressions](#index-expressions)
2410 (`expr[expr]`), and [field references](#field-expressions) (`expr.f`).
2411 All other expressions are rvalues.
2413 The left operand of an [assignment](#assignment-expressions) or
2414 [compound-assignment](#compound-assignment-expressions) expression is an lvalue context,
2415 as is the single operand of a unary [borrow](#unary-operator-expressions).
2416 All other expression contexts are rvalue contexts.
2418 When an lvalue is evaluated in an _lvalue context_, it denotes a memory location;
2419 when evaluated in an _rvalue context_, it denotes the value held _in_ that memory location.
2421 When an rvalue is used in lvalue context, a temporary un-named lvalue is created and used instead.
2422 A temporary's lifetime equals the largest lifetime of any reference that points to it.
2424 #### Moved and copied types
2426 When a [local variable](#memory-slots) is used
2427 as an [rvalue](#lvalues-rvalues-and-temporaries)
2428 the variable will either be moved or copied, depending on its type.
2429 For types that contain [owning pointers](#pointer-types)
2430 or values that implement the special trait `Drop`,
2431 the variable is moved.
2432 All other types are copied.
2434 ### Literal expressions
2436 A _literal expression_ consists of one of the [literal](#literals)
2437 forms described earlier. It directly describes a number, character,
2438 string, boolean value, or the unit value.
2442 "hello"; // string type
2443 '5'; // character type
2447 ### Path expressions
2449 A [path](#paths) used as an expression context denotes either a local variable or an item.
2450 Path expressions are [lvalues](#lvalues-rvalues-and-temporaries).
2452 ### Tuple expressions
2454 Tuples are written by enclosing one or more comma-separated
2455 expressions in parentheses. They are used to create [tuple-typed](#tuple-types)
2464 ### Structure expressions
2466 ~~~~ {.notrust .ebnf .gram}
2467 struct_expr : expr_path '{' ident ':' expr
2468 [ ',' ident ':' expr ] *
2471 [ ',' expr ] * ')' |
2475 There are several forms of structure expressions.
2476 A _structure expression_ consists of the [path](#paths) of a [structure item](#structures),
2477 followed by a brace-enclosed list of one or more comma-separated name-value pairs,
2478 providing the field values of a new instance of the structure.
2479 A field name can be any identifier, and is separated from its value expression by a colon.
2480 The location denoted by a structure field is mutable if and only if the enclosing structure is mutable.
2482 A _tuple structure expression_ consists of the [path](#paths) of a [structure item](#structures),
2483 followed by a parenthesized list of one or more comma-separated expressions
2484 (in other words, the path of a structure item followed by a tuple expression).
2485 The structure item must be a tuple structure item.
2487 A _unit-like structure expression_ consists only of the [path](#paths) of a [structure item](#structures).
2489 The following are examples of structure expressions:
2492 # struct Point { x: f64, y: f64 }
2493 # struct TuplePoint(f64, f64);
2494 # mod game { pub struct User<'a> { pub name: &'a str, pub age: uint, pub score: uint } }
2495 # struct Cookie; fn some_fn<T>(t: T) {}
2496 Point {x: 10.0, y: 20.0};
2497 TuplePoint(10.0, 20.0);
2498 let u = game::User {name: "Joe", age: 35, score: 100_000};
2499 some_fn::<Cookie>(Cookie);
2502 A structure expression forms a new value of the named structure type.
2503 Note that for a given *unit-like* structure type, this will always be the same value.
2505 A structure expression can terminate with the syntax `..` followed by an expression to denote a functional update.
2506 The expression following `..` (the base) must have the same structure type as the new structure type being formed.
2507 The entire expression denotes the result of allocating a new structure
2508 (with the same type as the base expression)
2509 with the given values for the fields that were explicitly specified
2510 and the values in the base record for all other fields.
2513 # struct Point3d { x: int, y: int, z: int }
2514 let base = Point3d {x: 1, y: 2, z: 3};
2515 Point3d {y: 0, z: 10, .. base};
2518 ### Block expressions
2520 ~~~~ {.notrust .ebnf .gram}
2521 block_expr : '{' [ view_item ] *
2522 [ stmt ';' | item ] *
2526 A _block expression_ is similar to a module in terms of the declarations that
2527 are possible. Each block conceptually introduces a new namespace scope. View
2528 items can bring new names into scopes and declared items are in scope for only
2531 A block will execute each statement sequentially, and then execute the
2532 expression (if given). If the final expression is omitted, the type and return
2533 value of the block are `()`, but if it is provided, the type and return value
2534 of the block are that of the expression itself.
2536 ### Method-call expressions
2538 ~~~~ {.notrust .ebnf .gram}
2539 method_call_expr : expr '.' ident paren_expr_list ;
2542 A _method call_ consists of an expression followed by a single dot, an identifier, and a parenthesized expression-list.
2543 Method calls are resolved to methods on specific traits,
2544 either statically dispatching to a method if the exact `self`-type of the left-hand-side is known,
2545 or dynamically dispatching if the left-hand-side expression is an indirect [object type](#object-types).
2547 ### Field expressions
2549 ~~~~ {.notrust .ebnf .gram}
2550 field_expr : expr '.' ident ;
2553 A _field expression_ consists of an expression followed by a single dot and an identifier,
2554 when not immediately followed by a parenthesized expression-list (the latter is a [method call expression](#method-call-expressions)).
2555 A field expression denotes a field of a [structure](#structure-types).
2557 ~~~~ {.ignore .field}
2560 (Struct {a: 10, b: 20}).a;
2563 A field access on a record is an [lvalue](#lvalues-rvalues-and-temporaries) referring to the value of that field.
2564 When the field is mutable, it can be [assigned](#assignment-expressions) to.
2566 When the type of the expression to the left of the dot is a pointer to a record or structure,
2567 it is automatically dereferenced to make the field access possible.
2569 ### Vector expressions
2571 ~~~~ {.notrust .ebnf .gram}
2572 vec_expr : '[' "mut" ? vec_elems? ']' ;
2574 vec_elems : [expr [',' expr]*] | [expr ',' ".." expr] ;
2577 A [_vector_](#vector-types) _expression_ is written by enclosing zero or
2578 more comma-separated expressions of uniform type in square brackets.
2580 In the `[expr ',' ".." expr]` form, the expression after the `".."`
2581 must be a constant expression that can be evaluated at compile time, such
2582 as a [literal](#literals) or a [static item](#static-items).
2586 ["a", "b", "c", "d"];
2587 [0, ..128]; // vector with 128 zeros
2588 [0u8, 0u8, 0u8, 0u8];
2591 ### Index expressions
2593 ~~~~ {.notrust .ebnf .gram}
2594 idx_expr : expr '[' expr ']' ;
2597 [Vector](#vector-types)-typed expressions can be indexed by writing a
2598 square-bracket-enclosed expression (the index) after them. When the
2599 vector is mutable, the resulting [lvalue](#lvalues-rvalues-and-temporaries) can be assigned to.
2601 Indices are zero-based, and may be of any integral type. Vector access
2602 is bounds-checked at run-time. When the check fails, it will put the
2603 task in a _failing state_.
2607 # task::spawn(proc() {
2610 (["a", "b"])[10]; // fails
2615 ### Unary operator expressions
2617 Rust defines six symbolic unary operators.
2618 They are all written as prefix operators,
2619 before the expression they apply to.
2622 : Negation. May only be applied to numeric types.
2624 : Dereference. When applied to a [pointer](#pointer-types) it denotes the pointed-to location.
2625 For pointers to mutable locations, the resulting [lvalue](#lvalues-rvalues-and-temporaries) can be assigned to.
2626 On non-pointer types, it calls the `deref` method of the `std::ops::Deref` trait, or the
2627 `deref_mut` method of the `std::ops::DerefMut` trait (if implemented by the type and required
2628 for an outer expression that will or could mutate the dereference), and produces the
2629 result of dereferencing the `&` or `&mut` borrowed pointer returned from the overload method.
2632 : Logical negation. On the boolean type, this flips between `true` and
2633 `false`. On integer types, this inverts the individual bits in the
2634 two's complement representation of the value.
2636 : [Boxing](#pointer-types) operators. Allocate a box to hold the value they are applied to,
2637 and store the value in it. `~` creates an owned box.
2639 : Borrow operator. Returns a reference, pointing to its operand.
2640 The operand of a borrow is statically proven to outlive the resulting pointer.
2641 If the borrow-checker cannot prove this, it is a compilation error.
2643 ### Binary operator expressions
2645 ~~~~ {.notrust .ebnf .gram}
2646 binop_expr : expr binop expr ;
2649 Binary operators expressions are given in terms of
2650 [operator precedence](#operator-precedence).
2652 #### Arithmetic operators
2654 Binary arithmetic expressions are syntactic sugar for calls to built-in traits,
2655 defined in the `std::ops` module of the `std` library.
2656 This means that arithmetic operators can be overridden for user-defined types.
2657 The default meaning of the operators on standard types is given here.
2660 : Addition and vector/string concatenation.
2661 Calls the `add` method on the `std::ops::Add` trait.
2664 Calls the `sub` method on the `std::ops::Sub` trait.
2667 Calls the `mul` method on the `std::ops::Mul` trait.
2670 Calls the `div` method on the `std::ops::Div` trait.
2673 Calls the `rem` method on the `std::ops::Rem` trait.
2675 #### Bitwise operators
2677 Like the [arithmetic operators](#arithmetic-operators), bitwise operators
2678 are syntactic sugar for calls to methods of built-in traits.
2679 This means that bitwise operators can be overridden for user-defined types.
2680 The default meaning of the operators on standard types is given here.
2684 Calls the `bitand` method of the `std::ops::BitAnd` trait.
2687 Calls the `bitor` method of the `std::ops::BitOr` trait.
2690 Calls the `bitxor` method of the `std::ops::BitXor` trait.
2692 : Logical left shift.
2693 Calls the `shl` method of the `std::ops::Shl` trait.
2695 : Logical right shift.
2696 Calls the `shr` method of the `std::ops::Shr` trait.
2698 #### Lazy boolean operators
2700 The operators `||` and `&&` may be applied to operands of boolean type.
2701 The `||` operator denotes logical 'or', and the `&&` operator denotes logical 'and'.
2702 They differ from `|` and `&` in that the right-hand operand is only evaluated
2703 when the left-hand operand does not already determine the result of the expression.
2704 That is, `||` only evaluates its right-hand operand
2705 when the left-hand operand evaluates to `false`, and `&&` only when it evaluates to `true`.
2707 #### Comparison operators
2709 Comparison operators are, like the [arithmetic operators](#arithmetic-operators),
2710 and [bitwise operators](#bitwise-operators),
2711 syntactic sugar for calls to built-in traits.
2712 This means that comparison operators can be overridden for user-defined types.
2713 The default meaning of the operators on standard types is given here.
2717 Calls the `eq` method on the `std::cmp::Eq` trait.
2720 Calls the `ne` method on the `std::cmp::Eq` trait.
2723 Calls the `lt` method on the `std::cmp::Ord` trait.
2726 Calls the `gt` method on the `std::cmp::Ord` trait.
2728 : Less than or equal.
2729 Calls the `le` method on the `std::cmp::Ord` trait.
2731 : Greater than or equal.
2732 Calls the `ge` method on the `std::cmp::Ord` trait.
2734 #### Type cast expressions
2736 A type cast expression is denoted with the binary operator `as`.
2738 Executing an `as` expression casts the value on the left-hand side to the type
2739 on the right-hand side.
2741 A numeric value can be cast to any numeric type.
2742 A raw pointer value can be cast to or from any integral type or raw pointer type.
2743 Any other cast is unsupported and will fail to compile.
2745 An example of an `as` expression:
2748 # fn sum(v: &[f64]) -> f64 { 0.0 }
2749 # fn len(v: &[f64]) -> int { 0 }
2751 fn avg(v: &[f64]) -> f64 {
2752 let sum: f64 = sum(v);
2753 let sz: f64 = len(v) as f64;
2758 #### Assignment expressions
2760 An _assignment expression_ consists of an [lvalue](#lvalues-rvalues-and-temporaries) expression followed by an
2761 equals sign (`=`) and an [rvalue](#lvalues-rvalues-and-temporaries) expression.
2763 Evaluating an assignment expression [either copies or moves](#moved-and-copied-types) its right-hand operand to its left-hand operand.
2772 #### Compound assignment expressions
2774 The `+`, `-`, `*`, `/`, `%`, `&`, `|`, `^`, `<<`, and `>>`
2775 operators may be composed with the `=` operator. The expression `lval
2776 OP= val` is equivalent to `lval = lval OP val`. For example, `x = x +
2777 1` may be written as `x += 1`.
2779 Any such expression always has the [`unit`](#primitive-types) type.
2781 #### Operator precedence
2783 The precedence of Rust binary operators is ordered as follows, going
2784 from strong to weak:
2786 ~~~~ {.notrust .precedence}
2801 Operators at the same precedence level are evaluated left-to-right. [Unary operators](#unary-operator-expressions)
2802 have the same precedence level and it is stronger than any of the binary operators'.
2804 ### Grouped expressions
2806 An expression enclosed in parentheses evaluates to the result of the enclosed
2807 expression. Parentheses can be used to explicitly specify evaluation order
2808 within an expression.
2810 ~~~~ {.notrust .ebnf .gram}
2811 paren_expr : '(' expr ')' ;
2814 An example of a parenthesized expression:
2817 let x = (2 + 3) * 4;
2821 ### Call expressions
2823 ~~~~ {.notrust .ebnf .gram}
2824 expr_list : [ expr [ ',' expr ]* ] ? ;
2825 paren_expr_list : '(' expr_list ')' ;
2826 call_expr : expr paren_expr_list ;
2829 A _call expression_ invokes a function, providing zero or more input slots and
2830 an optional reference slot to serve as the function's output, bound to the
2831 `lval` on the right hand side of the call. If the function eventually returns,
2832 then the expression completes.
2834 Some examples of call expressions:
2837 # use std::from_str::FromStr;
2838 # fn add(x: int, y: int) -> int { 0 }
2840 let x: int = add(1, 2);
2841 let pi: Option<f32> = FromStr::from_str("3.14");
2844 ### Lambda expressions
2846 ~~~~ {.notrust .ebnf .gram}
2847 ident_list : [ ident [ ',' ident ]* ] ? ;
2848 lambda_expr : '|' ident_list '|' expr ;
2851 A _lambda expression_ (sometimes called an "anonymous function expression") defines a function and denotes it as a value,
2852 in a single expression.
2853 A lambda expression is a pipe-symbol-delimited (`|`) list of identifiers followed by an expression.
2855 A lambda expression denotes a function that maps a list of parameters (`ident_list`)
2856 onto the expression that follows the `ident_list`.
2857 The identifiers in the `ident_list` are the parameters to the function.
2858 These parameters' types need not be specified, as the compiler infers them from context.
2860 Lambda expressions are most useful when passing functions as arguments to other functions,
2861 as an abbreviation for defining and capturing a separate function.
2863 Significantly, lambda expressions _capture their environment_,
2864 which regular [function definitions](#functions) do not.
2865 The exact type of capture depends on the [function type](#function-types) inferred for the lambda expression.
2866 In the simplest and least-expensive form (analogous to a ```|| { }``` expression),
2867 the lambda expression captures its environment by reference,
2868 effectively borrowing pointers to all outer variables mentioned inside the function.
2869 Alternately, the compiler may infer that a lambda expression should copy or move values (depending on their type.)
2870 from the environment into the lambda expression's captured environment.
2872 In this example, we define a function `ten_times` that takes a higher-order function argument,
2873 and call it with a lambda expression as an argument.
2876 fn ten_times(f: |int|) {
2884 ten_times(|j| println!("hello, {}", j));
2889 ~~~~ {.notrust .ebnf .gram}
2890 while_expr : "while" expr '{' block '}' ;
2893 A `while` loop begins by evaluating the boolean loop conditional expression.
2894 If the loop conditional expression evaluates to `true`, the loop body block
2895 executes and control returns to the loop conditional expression. If the loop
2896 conditional expression evaluates to `false`, the `while` expression completes.
2911 The keyword `loop` in Rust appears both in _loop expressions_ and in _continue expressions_.
2912 A loop expression denotes an infinite loop;
2913 see [Continue expressions](#continue-expressions) for continue expressions.
2915 ~~~~ {.notrust .ebnf .gram}
2916 loop_expr : [ lifetime ':' ] "loop" '{' block '}';
2919 A `loop` expression may optionally have a _label_.
2920 If a label is present,
2921 then labeled `break` and `loop` expressions nested within this loop may exit out of this loop or return control to its head.
2922 See [Break expressions](#break-expressions).
2924 ### Break expressions
2926 ~~~~ {.notrust .ebnf .gram}
2927 break_expr : "break" [ lifetime ];
2930 A `break` expression has an optional `label`.
2931 If the label is absent, then executing a `break` expression immediately terminates the innermost loop enclosing it.
2932 It is only permitted in the body of a loop.
2933 If the label is present, then `break foo` terminates the loop with label `foo`,
2934 which need not be the innermost label enclosing the `break` expression,
2935 but must enclose it.
2937 ### Continue expressions
2939 ~~~~ {.notrust .ebnf .gram}
2940 continue_expr : "loop" [ lifetime ];
2943 A continue expression, written `loop`, also has an optional `label`.
2944 If the label is absent,
2945 then executing a `loop` expression immediately terminates the current iteration of the innermost loop enclosing it,
2946 returning control to the loop *head*.
2947 In the case of a `while` loop,
2948 the head is the conditional expression controlling the loop.
2949 In the case of a `for` loop, the head is the call-expression controlling the loop.
2950 If the label is present, then `loop foo` returns control to the head of the loop with label `foo`,
2951 which need not be the innermost label enclosing the `break` expression,
2952 but must enclose it.
2954 A `loop` expression is only permitted in the body of a loop.
2958 ~~~~ {.notrust .ebnf .gram}
2959 for_expr : "for" pat "in" expr '{' block '}' ;
2962 A `for` expression is a syntactic construct for looping over elements
2963 provided by an implementation of `std::iter::Iterator`.
2965 An example of a for loop over the contents of a vector:
2969 # fn bar(f: Foo) { }
2974 let v: &[Foo] = &[a, b, c];
2981 An example of a for loop over a series of integers:
2984 # fn bar(b:uint) { }
2985 for i in range(0u, 256) {
2992 ~~~~ {.notrust .ebnf .gram}
2993 if_expr : "if" expr '{' block '}'
2996 else_tail : "else" [ if_expr
3000 An `if` expression is a conditional branch in program control. The form of
3001 an `if` expression is a condition expression, followed by a consequent
3002 block, any number of `else if` conditions and blocks, and an optional
3003 trailing `else` block. The condition expressions must have type
3004 `bool`. If a condition expression evaluates to `true`, the
3005 consequent block is executed and any subsequent `else if` or `else`
3006 block is skipped. If a condition expression evaluates to `false`, the
3007 consequent block is skipped and any subsequent `else if` condition is
3008 evaluated. If all `if` and `else if` conditions evaluate to `false`
3009 then any `else` block is executed.
3011 ### Match expressions
3013 ~~~~ {.notrust .ebnf .gram}
3014 match_expr : "match" expr '{' match_arm [ '|' match_arm ] * '}' ;
3016 match_arm : match_pat "=>" [ expr "," | '{' block '}' ] ;
3018 match_pat : pat [ ".." pat ] ? [ "if" expr ] ;
3021 A `match` expression branches on a *pattern*. The exact form of matching that
3022 occurs depends on the pattern. Patterns consist of some combination of
3023 literals, destructured vectors or enum constructors, structures, records and
3024 tuples, variable binding specifications, wildcards (`..`), and placeholders
3025 (`_`). A `match` expression has a *head expression*, which is the value to
3026 compare to the patterns. The type of the patterns must equal the type of the
3029 In a pattern whose head expression has an `enum` type, a placeholder (`_`)
3030 stands for a *single* data field, whereas a wildcard `..` stands for *all* the
3031 fields of a particular variant. For example:
3034 enum List<X> { Nil, Cons(X, ~List<X>) }
3036 let x: List<int> = Cons(10, ~Cons(11, ~Nil));
3039 Cons(_, ~Nil) => fail!("singleton list"),
3041 Nil => fail!("empty list")
3045 The first pattern matches lists constructed by applying `Cons` to any head
3046 value, and a tail value of `~Nil`. The second pattern matches _any_ list
3047 constructed with `Cons`, ignoring the values of its arguments. The difference
3048 between `_` and `..` is that the pattern `C(_)` is only type-correct if `C` has
3049 exactly one argument, while the pattern `C(..)` is type-correct for any enum
3050 variant `C`, regardless of how many arguments `C` has.
3052 Used inside a vector pattern, `..` stands for any number of elements. This
3053 wildcard can be used at most once for a given vector, which implies that it
3054 cannot be used to specifically match elements that are at an unknown distance
3055 from both ends of a vector, like `[.., 42, ..]`. If followed by a variable name,
3056 it will bind the corresponding slice to the variable. Example:
3059 fn is_symmetric(list: &[uint]) -> bool {
3062 [x, ..inside, y] if x == y => is_symmetric(inside),
3068 let sym = &[0, 1, 4, 2, 4, 1, 0];
3069 let not_sym = &[0, 1, 7, 2, 4, 1, 0];
3070 assert!(is_symmetric(sym));
3071 assert!(!is_symmetric(not_sym));
3075 A `match` behaves differently depending on whether or not the head expression
3076 is an [lvalue or an rvalue](#lvalues-rvalues-and-temporaries).
3077 If the head expression is an rvalue, it is
3078 first evaluated into a temporary location, and the resulting value
3079 is sequentially compared to the patterns in the arms until a match
3080 is found. The first arm with a matching pattern is chosen as the branch target
3081 of the `match`, any variables bound by the pattern are assigned to local
3082 variables in the arm's block, and control enters the block.
3084 When the head expression is an lvalue, the match does not allocate a
3085 temporary location (however, a by-value binding may copy or move from
3086 the lvalue). When possible, it is preferable to match on lvalues, as the
3087 lifetime of these matches inherits the lifetime of the lvalue, rather
3088 than being restricted to the inside of the match.
3090 An example of a `match` expression:
3093 # fn process_pair(a: int, b: int) { }
3094 # fn process_ten() { }
3096 enum List<X> { Nil, Cons(X, ~List<X>) }
3098 let x: List<int> = Cons(10, ~Cons(11, ~Nil));
3101 Cons(a, ~Cons(b, _)) => {
3116 Patterns that bind variables
3117 default to binding to a copy or move of the matched value
3118 (depending on the matched value's type).
3119 This can be changed to bind to a reference by
3120 using the `ref` keyword,
3121 or to a mutable reference using `ref mut`.
3123 Subpatterns can also be bound to variables by the use of the syntax
3124 `variable @ pattern`.
3128 enum List { Nil, Cons(uint, ~List) }
3130 fn is_sorted(list: &List) -> bool {
3132 Nil | Cons(_, ~Nil) => true,
3133 Cons(x, ref r @ ~Cons(y, _)) => (x <= y) && is_sorted(*r)
3138 let a = Cons(6, ~Cons(7, ~Cons(42, ~Nil)));
3139 assert!(is_sorted(&a));
3144 Patterns can also dereference pointers by using the `&`,
3145 `~` or `@` symbols, as appropriate. For example, these two matches
3146 on `x: &int` are equivalent:
3150 let y = match *x { 0 => "zero", _ => "some" };
3151 let z = match x { &0 => "zero", _ => "some" };
3156 A pattern that's just an identifier, like `Nil` in the previous example,
3157 could either refer to an enum variant that's in scope, or bind a new variable.
3158 The compiler resolves this ambiguity by forbidding variable bindings that occur
3159 in `match` patterns from shadowing names of variants that are in scope.
3160 For example, wherever `List` is in scope,
3161 a `match` pattern would not be able to bind `Nil` as a new name.
3162 The compiler interprets a variable pattern `x` as a binding _only_ if there is
3163 no variant named `x` in scope.
3164 A convention you can use to avoid conflicts is simply to name variants with
3165 upper-case letters, and local variables with lower-case letters.
3167 Multiple match patterns may be joined with the `|` operator.
3168 A range of values may be specified with `..`.
3174 let message = match x {
3175 0 | 1 => "not many",
3181 Range patterns only work on scalar types
3182 (like integers and characters; not like vectors and structs, which have sub-components).
3183 A range pattern may not be a sub-range of another range pattern inside the same `match`.
3185 Finally, match patterns can accept *pattern guards* to further refine the
3186 criteria for matching a case. Pattern guards appear after the pattern and
3187 consist of a bool-typed expression following the `if` keyword. A pattern
3188 guard may refer to the variables bound within the pattern they follow.
3191 # let maybe_digit = Some(0);
3192 # fn process_digit(i: int) { }
3193 # fn process_other(i: int) { }
3195 let message = match maybe_digit {
3196 Some(x) if x < 10 => process_digit(x),
3197 Some(x) => process_other(x),
3202 ### Return expressions
3204 ~~~~ {.notrust .ebnf .gram}
3205 return_expr : "return" expr ? ;
3208 Return expressions are denoted with the keyword `return`. Evaluating a `return`
3209 expression moves its argument into the output slot of the current
3210 function, destroys the current function activation frame, and transfers
3211 control to the caller frame.
3213 An example of a `return` expression:
3216 fn max(a: int, b: int) -> int {
3228 Every slot, item and value in a Rust program has a type. The _type_ of a *value*
3229 defines the interpretation of the memory holding it.
3231 Built-in types and type-constructors are tightly integrated into the language,
3232 in nontrivial ways that are not possible to emulate in user-defined
3233 types. User-defined types have limited capabilities.
3237 The primitive types are the following:
3239 * The "unit" type `()`, having the single "unit" value `()` (occasionally called "nil").
3240 ^[The "unit" value `()` is *not* a sentinel "null pointer" value for reference slots; the "unit" type is the implicit return type from functions otherwise lacking a return type, and can be used in other contexts (such as message-sending or type-parametric code) as a zero-size type.]
3241 * The boolean type `bool` with values `true` and `false`.
3242 * The machine types.
3243 * The machine-dependent integer and floating-point types.
3247 The machine types are the following:
3249 * The unsigned word types `u8`, `u16`, `u32` and `u64`, with values drawn from
3250 the integer intervals $[0, 2^8 - 1]$, $[0, 2^{16} - 1]$, $[0, 2^{32} - 1]$ and
3251 $[0, 2^{64} - 1]$ respectively.
3253 * The signed two's complement word types `i8`, `i16`, `i32` and `i64`, with
3254 values drawn from the integer intervals $[-(2^7), 2^7 - 1]$,
3255 $[-(2^{15}), 2^{15} - 1]$, $[-(2^{31}), 2^{31} - 1]$, $[-(2^{63}), 2^{63} - 1]$
3258 * The IEEE 754-2008 `binary32` and `binary64` floating-point types: `f32` and
3259 `f64`, respectively.
3261 #### Machine-dependent integer types
3263 The Rust type `uint`^[A Rust `uint` is analogous to a C99 `uintptr_t`.] is an
3264 unsigned integer type with target-machine-dependent size. Its size, in
3265 bits, is equal to the number of bits required to hold any memory address on
3268 The Rust type `int`^[A Rust `int` is analogous to a C99 `intptr_t`.] is a
3269 two's complement signed integer type with target-machine-dependent size. Its
3270 size, in bits, is equal to the size of the rust type `uint` on the same target
3275 The types `char` and `str` hold textual data.
3277 A value of type `char` is a [Unicode scalar value](
3278 http://www.unicode.org/glossary/#unicode_scalar_value)
3279 (ie. a code point that is not a surrogate),
3280 represented as a 32-bit unsigned word in the 0x0000 to 0xD7FF
3281 or 0xE000 to 0x10FFFF range.
3282 A `[char]` vector is effectively an UCS-4 / UTF-32 string.
3284 A value of type `str` is a Unicode string,
3285 represented as a vector of 8-bit unsigned bytes holding a sequence of UTF-8 codepoints.
3286 Since `str` is of unknown size, it is not a _first class_ type,
3287 but can only be instantiated through a pointer type,
3288 such as `&str` or `~str`.
3292 The tuple type-constructor forms a new heterogeneous product of values similar
3293 to the record type-constructor. The differences are as follows:
3295 * tuple elements cannot be mutable, unlike record fields
3296 * tuple elements are not named and can be accessed only by pattern-matching
3298 Tuple types and values are denoted by listing the types or values of their
3299 elements, respectively, in a parenthesized, comma-separated
3302 The members of a tuple are laid out in memory contiguously, like a record, in
3303 order specified by the tuple type.
3305 An example of a tuple type and its use:
3308 type Pair<'a> = (int,&'a str);
3309 let p: Pair<'static> = (10,"hello");
3311 assert!(b != "world");
3316 The vector type constructor represents a homogeneous array of values of a given type.
3317 A vector has a fixed size.
3318 (Operations like `vec.push` operate solely on owned vectors.)
3319 A vector type can be annotated with a _definite_ size, such as `[int, ..10]`.
3320 Such a definite-sized vector type is a first-class type, since its size is known statically.
3321 A vector without such a size is said to be of _indefinite_ size,
3322 and is therefore not a _first-class_ type.
3323 An indefinite-size vector can only be instantiated through a pointer type,
3324 such as `&[T]` or `~[T]`.
3325 The kind of a vector type depends on the kind of its element type,
3326 as with other simple structural types.
3328 Expressions producing vectors of definite size cannot be evaluated in a
3329 context expecting a vector of indefinite size; one must copy the
3330 definite-sized vector contents into a distinct vector of indefinite size.
3332 An example of a vector type and its use:
3335 let v: &[int] = &[7, 5, 3];
3340 All in-bounds elements of a vector are always initialized,
3341 and access to a vector is always bounds-checked.
3345 A `struct` *type* is a heterogeneous product of other types, called the *fields* of the type.
3346 ^[`struct` types are analogous `struct` types in C,
3347 the *record* types of the ML family,
3348 or the *structure* types of the Lisp family.]
3350 New instances of a `struct` can be constructed with a [struct expression](#structure-expressions).
3352 The memory order of fields in a `struct` is given by the item defining it.
3353 Fields may be given in any order in a corresponding struct *expression*;
3354 the resulting `struct` value will always be laid out in memory in the order specified by the corresponding *item*.
3356 The fields of a `struct` may be qualified by [visibility modifiers](#re-exporting-and-visibility),
3357 to restrict access to implementation-private data in a structure.
3359 A _tuple struct_ type is just like a structure type, except that the fields are anonymous.
3361 A _unit-like struct_ type is like a structure type, except that it has no fields.
3362 The one value constructed by the associated [structure expression](#structure-expressions)
3363 is the only value that inhabits such a type.
3365 ### Enumerated types
3367 An *enumerated type* is a nominal, heterogeneous disjoint union type,
3368 denoted by the name of an [`enum` item](#enumerations).
3369 ^[The `enum` type is analogous to a `data` constructor declaration in ML,
3370 or a *pick ADT* in Limbo.]
3372 An [`enum` item](#enumerations) declares both the type and a number of *variant constructors*,
3373 each of which is independently named and takes an optional tuple of arguments.
3375 New instances of an `enum` can be constructed by calling one of the variant constructors,
3376 in a [call expression](#call-expressions).
3378 Any `enum` value consumes as much memory as the largest variant constructor for its corresponding `enum` type.
3380 Enum types cannot be denoted *structurally* as types,
3381 but must be denoted by named reference to an [`enum` item](#enumerations).
3385 Nominal types -- [enumerations](#enumerated-types) and [structures](#structure-types) -- may be recursive.
3386 That is, each `enum` constructor or `struct` field may refer, directly or indirectly, to the enclosing `enum` or `struct` type itself.
3387 Such recursion has restrictions:
3389 * Recursive types must include a nominal type in the recursion
3390 (not mere [type definitions](#type-definitions),
3391 or other structural types such as [vectors](#vector-types) or [tuples](#tuple-types)).
3392 * A recursive `enum` item must have at least one non-recursive constructor
3393 (in order to give the recursion a basis case).
3394 * The size of a recursive type must be finite;
3395 in other words the recursive fields of the type must be [pointer types](#pointer-types).
3396 * Recursive type definitions can cross module boundaries, but not module *visibility* boundaries,
3397 or crate boundaries (in order to simplify the module system and type checker).
3399 An example of a *recursive* type and its use:
3407 let a: List<int> = Cons(7, ~Cons(13, ~Nil));
3412 All pointers in Rust are explicit first-class values.
3413 They can be copied, stored into data structures, and returned from functions.
3414 There are four varieties of pointer in Rust:
3416 * Owning pointers (`~`)
3417 : These point to owned heap allocations (or "boxes") in the shared, inter-task heap.
3418 Each owned box has a single owning pointer; pointer and pointee retain a 1:1 relationship at all times.
3419 Owning pointers are written `~content`,
3420 for example `~int` means an owning pointer to an owned box containing an integer.
3421 Copying an owned box is a "deep" operation:
3422 it involves allocating a new owned box and copying the contents of the old box into the new box.
3423 Releasing an owning pointer immediately releases its corresponding owned box.
3426 : These point to memory _owned by some other value_.
3427 References arise by (automatic) conversion from owning pointers, managed pointers,
3428 or by applying the borrowing operator `&` to some other value,
3429 including [lvalues, rvalues or temporaries](#lvalues-rvalues-and-temporaries).
3430 References are written `&content`, or in some cases `&'f content` for some lifetime-variable `f`,
3431 for example `&int` means a reference to an integer.
3432 Copying a reference is a "shallow" operation:
3433 it involves only copying the pointer itself.
3434 Releasing a reference typically has no effect on the value it points to,
3435 with the exception of temporary values,
3436 which are released when the last reference to them is released.
3438 * Raw pointers (`*`)
3439 : Raw pointers are pointers without safety or liveness guarantees.
3440 Raw pointers are written `*content`,
3441 for example `*int` means a raw pointer to an integer.
3442 Copying or dropping a raw pointer has no effect on the lifecycle of any other value.
3443 Dereferencing a raw pointer or converting it to any other pointer type is an [`unsafe` operation](#unsafe-functions).
3444 Raw pointers are generally discouraged in Rust code;
3445 they exist to support interoperability with foreign code,
3446 and writing performance-critical or low-level functions.
3450 The function type constructor `fn` forms new function types.
3451 A function type consists of a possibly-empty set of function-type modifiers
3452 (such as `unsafe` or `extern`), a sequence of input types and an output type.
3454 An example of a `fn` type:
3457 fn add(x: int, y: int) -> int {
3461 let mut x = add(5,7);
3463 type Binop<'a> = |int,int|: 'a -> int;
3464 let bo: Binop = add;
3470 ~~~~ {.notrust .ebnf .notation}
3471 closure_type := [ 'unsafe' ] [ '<' lifetime-list '>' ] '|' arg-list '|'
3472 [ ':' bound-list ] [ '->' type ]
3473 procedure_type := 'proc' [ '<' lifetime-list '>' ] '(' arg-list ')'
3474 [ ':' bound-list ] [ '->' type ]
3475 lifetime-list := lifetime | lifetime ',' lifetime-list
3476 arg-list := ident ':' type | ident ':' type ',' arg-list
3477 bound-list := bound | bound '+' bound-list
3478 bound := path | lifetime
3481 The type of a closure mapping an input of type `A` to an output of type `B` is
3482 `|A| -> B`. A closure with no arguments or return values has type `||`.
3483 Similarly, a procedure mapping `A` to `B` is `proc(A) -> B` and a no-argument
3484 and no-return value closure has type `proc()`.
3486 An example of creating and calling a closure:
3489 let captured_var = 10;
3491 let closure_no_args = || println!("captured_var={}", captured_var);
3493 let closure_args = |arg: int| -> int {
3494 println!("captured_var={}, arg={}", captured_var, arg);
3495 arg // Note lack of semicolon after 'arg'
3498 fn call_closure(c1: ||, c2: |int| -> int) {
3503 call_closure(closure_no_args, closure_args);
3507 Unlike closures, procedures may only be invoked once, but own their
3508 environment, and are allowed to move out of their environment. Procedures are
3509 allocated on the heap (unlike closures). An example of creating and calling a
3513 let string = ~"Hello";
3515 // Creates a new procedure, passing it to the `spawn` function.
3517 println!("{} world!", string);
3520 // the variable `string` has been moved into the previous procedure, so it is
3521 // no longer usable.
3524 // Create an invoke a procedure. Note that the procedure is *moved* when
3525 // invoked, so it cannot be invoked again.
3526 let f = proc(n: int) { n + 22 };
3527 println!("answer: {}", f(20));
3533 Every trait item (see [traits](#traits)) defines a type with the same name as the trait.
3534 This type is called the _object type_ of the trait.
3535 Object types permit "late binding" of methods, dispatched using _virtual method tables_ ("vtables").
3536 Whereas most calls to trait methods are "early bound" (statically resolved) to specific implementations at compile time,
3537 a call to a method on an object type is only resolved to a vtable entry at compile time.
3538 The actual implementation for each vtable entry can vary on an object-by-object basis.
3540 Given a pointer-typed expression `E` of type `&T` or `~T`, where `T` implements trait `R`,
3541 casting `E` to the corresponding pointer type `&R` or `~R` results in a value of the _object type_ `R`.
3542 This result is represented as a pair of pointers:
3543 the vtable pointer for the `T` implementation of `R`, and the pointer value of `E`.
3545 An example of an object type:
3549 fn to_string(&self) -> ~str;
3552 impl Printable for int {
3553 fn to_string(&self) -> ~str { self.to_str() }
3556 fn print(a: ~Printable) {
3557 println!("{}", a.to_string());
3561 print(~10 as ~Printable);
3565 In this example, the trait `Printable` occurs as an object type in both the type signature of `print`,
3566 and the cast expression in `main`.
3570 Within the body of an item that has type parameter declarations, the names of its type parameters are types:
3573 fn map<A: Clone, B: Clone>(f: |A| -> B, xs: &[A]) -> ~[B] {
3577 let first: B = f(xs[0].clone());
3578 let rest: ~[B] = map(f, xs.slice(1, xs.len()));
3579 return ~[first] + rest;
3583 Here, `first` has type `B`, referring to `map`'s `B` type parameter;
3584 and `rest` has type `~[B]`, a vector type with element type `B`.
3588 The special type `self` has a meaning within methods inside an
3589 impl item. It refers to the type of the implicit `self` argument. For
3594 fn make_string(&self) -> ~str;
3597 impl Printable for ~str {
3598 fn make_string(&self) -> ~str {
3604 `self` refers to the value of type `~str` that is the receiver for a
3605 call to the method `make_string`.
3609 Types in Rust are categorized into kinds, based on various properties of the components of the type.
3613 : Types of this kind can be safely sent between tasks.
3614 This kind includes scalars, owning pointers, owned closures, and
3615 structural types containing only other owned types.
3616 All `Send` types are `'static`.
3618 : Types of this kind consist of "Plain Old Data"
3619 which can be copied by simply moving bits.
3620 All values of this kind can be implicitly copied.
3621 This kind includes scalars and immutable references,
3622 as well as structural types containing other `Copy` types.
3624 : Types of this kind do not contain any references (except for
3625 references with the `static` lifetime, which are allowed).
3626 This can be a useful guarantee for code
3627 that breaks borrowing assumptions
3628 using [`unsafe` operations](#unsafe-functions).
3630 : This is not strictly a kind,
3631 but its presence interacts with kinds:
3632 the `Drop` trait provides a single method `drop`
3633 that takes no parameters,
3634 and is run when values of the type are dropped.
3635 Such a method is called a "destructor",
3636 and are always executed in "top-down" order:
3637 a value is completely destroyed
3638 before any of the values it owns run their destructors.
3639 Only `Send` types can implement `Drop`.
3642 : Types with destructors, closure environments,
3643 and various other _non-first-class_ types,
3644 are not copyable at all.
3645 Such types can usually only be accessed through pointers,
3646 or in some cases, moved between mutable locations.
3648 Kinds can be supplied as _bounds_ on type parameters, like traits,
3649 in which case the parameter is constrained to types satisfying that kind.
3651 By default, type parameters do not carry any assumed kind-bounds at all.
3652 When instantiating a type parameter,
3653 the kind bounds on the parameter are checked
3654 to be the same or narrower than the kind
3655 of the type that it is instantiated with.
3657 Sending operations are not part of the Rust language,
3658 but are implemented in the library.
3659 Generic functions that send values
3660 bound the kind of these values to sendable.
3662 # Memory and concurrency models
3664 Rust has a memory model centered around concurrently-executing _tasks_. Thus
3665 its memory model and its concurrency model are best discussed simultaneously,
3666 as parts of each only make sense when considered from the perspective of the
3669 When reading about the memory model, keep in mind that it is partitioned in
3670 order to support tasks; and when reading about tasks, keep in mind that their
3671 isolation and communication mechanisms are only possible due to the ownership
3672 and lifetime semantics of the memory model.
3676 A Rust program's memory consists of a static set of *items*, a set of
3677 [tasks](#tasks) each with its own *stack*, and a *heap*. Immutable portions of
3678 the heap may be shared between tasks, mutable portions may not.
3680 Allocations in the stack consist of *slots*, and allocations in the heap
3683 ### Memory allocation and lifetime
3685 The _items_ of a program are those functions, modules and types
3686 that have their value calculated at compile-time and stored uniquely in the
3687 memory image of the rust process. Items are neither dynamically allocated nor
3690 A task's _stack_ consists of activation frames automatically allocated on
3691 entry to each function as the task executes. A stack allocation is reclaimed
3692 when control leaves the frame containing it.
3694 The _heap_ is a general term that describes two separate sets of boxes:
3695 managed boxes -- which may be subject to garbage collection -- and owned
3696 boxes. The lifetime of an allocation in the heap depends on the lifetime of
3697 the box values pointing to it. Since box values may themselves be passed in
3698 and out of frames, or stored in the heap, heap allocations may outlive the
3699 frame they are allocated within.
3701 ### Memory ownership
3703 A task owns all memory it can *safely* reach through local variables,
3704 as well as managed, owned boxes and references.
3706 When a task sends a value that has the `Send` trait to another task,
3707 it loses ownership of the value sent and can no longer refer to it.
3708 This is statically guaranteed by the combined use of "move semantics",
3709 and the compiler-checked _meaning_ of the `Send` trait:
3710 it is only instantiated for (transitively) sendable kinds of data constructor and pointers,
3711 never including managed boxes or references.
3713 When a stack frame is exited, its local allocations are all released, and its
3714 references to boxes (both managed and owned) are dropped.
3716 A managed box may (in the case of a recursive, mutable managed type) be cyclic;
3717 in this case the release of memory inside the managed structure may be deferred
3718 until task-local garbage collection can reclaim it. Code can ensure no such
3719 delayed deallocation occurs by restricting itself to owned boxes and similar
3720 unmanaged kinds of data.
3722 When a task finishes, its stack is necessarily empty and it therefore has no
3723 references to any boxes; the remainder of its heap is immediately freed.
3727 A task's stack contains slots.
3729 A _slot_ is a component of a stack frame, either a function parameter,
3730 a [temporary](#lvalues-rvalues-and-temporaries), or a local variable.
3732 A _local variable_ (or *stack-local* allocation) holds a value directly,
3733 allocated within the stack's memory. The value is a part of the stack frame.
3735 Local variables are immutable unless declared otherwise like: `let mut x = ...`.
3737 Function parameters are immutable unless declared with `mut`. The
3738 `mut` keyword applies only to the following parameter (so `|mut x, y|`
3739 and `fn f(mut x: ~int, y: ~int)` declare one mutable variable `x` and
3740 one immutable variable `y`).
3742 Methods that take either `self` or `~self` can optionally place them in a
3743 mutable slot by prefixing them with `mut` (similar to regular arguments):
3747 fn change(mut self) -> Self;
3748 fn modify(mut ~self) -> ~Self;
3752 Local variables are not initialized when allocated; the entire frame worth of
3753 local variables are allocated at once, on frame-entry, in an uninitialized
3754 state. Subsequent statements within a function may or may not initialize the
3755 local variables. Local variables can be used only after they have been
3756 initialized; this is enforced by the compiler.
3760 An _owned box_ is a reference to a heap allocation holding another value, which is constructed
3761 by the prefix *tilde* sigil `~`
3763 An example of an owned box type and value:
3769 Owned box values exist in 1:1 correspondence with their heap allocation
3770 copying an owned box value makes a shallow copy of the pointer
3771 Rust will consider a shallow copy of an owned box to move ownership of the value. After a value has been moved, the source location cannot be used unless it is reinitialized.
3776 // attempting to use `x` will result in an error here
3783 An executing Rust program consists of a tree of tasks.
3784 A Rust _task_ consists of an entry function, a stack,
3785 a set of outgoing communication channels and incoming communication ports,
3786 and ownership of some portion of the heap of a single operating-system process.
3787 (We expect that many programs will not use channels and ports directly,
3788 but will instead use higher-level abstractions provided in standard libraries,
3791 Multiple Rust tasks may coexist in a single operating-system process.
3792 The runtime scheduler maps tasks to a certain number of operating-system threads.
3793 By default, the scheduler chooses the number of threads based on
3794 the number of concurrent physical CPUs detected at startup.
3795 It's also possible to override this choice at runtime.
3796 When the number of tasks exceeds the number of threads -- which is likely --
3797 the scheduler multiplexes the tasks onto threads.^[
3798 This is an M:N scheduler,
3799 which is known to give suboptimal results for CPU-bound concurrency problems.
3800 In such cases, running with the same number of threads and tasks can yield better results.
3801 Rust has M:N scheduling in order to support very large numbers of tasks
3802 in contexts where threads are too resource-intensive to use in large number.
3803 The cost of threads varies substantially per operating system, and is sometimes quite low,
3804 so this flexibility is not always worth exploiting.]
3806 ### Communication between tasks
3808 Rust tasks are isolated and generally unable to interfere with one another's memory directly,
3809 except through [`unsafe` code](#unsafe-functions).
3810 All contact between tasks is mediated by safe forms of ownership transfer,
3811 and data races on memory are prohibited by the type system.
3813 Inter-task communication and co-ordination facilities are provided in the standard library.
3816 - synchronous and asynchronous communication channels with various communication topologies
3817 - read-only and read-write shared variables with various safe mutual exclusion patterns
3818 - simple locks and semaphores
3820 When such facilities carry values, the values are restricted to the [`Send` type-kind](#type-kinds).
3821 Restricting communication interfaces to this kind ensures that no references or managed pointers move between tasks.
3822 Thus access to an entire data structure can be mediated through its owning "root" value;
3823 no further locking or copying is required to avoid data races within the substructure of such a value.
3827 The _lifecycle_ of a task consists of a finite set of states and events
3828 that cause transitions between the states. The lifecycle states of a task are:
3835 A task begins its lifecycle -- once it has been spawned -- in the *running*
3836 state. In this state it executes the statements of its entry function, and any
3837 functions called by the entry function.
3839 A task may transition from the *running* state to the *blocked*
3840 state any time it makes a blocking communication call. When the
3841 call can be completed -- when a message arrives at a sender, or a
3842 buffer opens to receive a message -- then the blocked task will
3843 unblock and transition back to *running*.
3845 A task may transition to the *failing* state at any time, due being
3846 killed by some external event or internally, from the evaluation of a
3847 `fail!()` macro. Once *failing*, a task unwinds its stack and
3848 transitions to the *dead* state. Unwinding the stack of a task is done by
3849 the task itself, on its own control stack. If a value with a destructor is
3850 freed during unwinding, the code for the destructor is run, also on the task's
3851 control stack. Running the destructor code causes a temporary transition to a
3852 *running* state, and allows the destructor code to cause any subsequent
3853 state transitions. The original task of unwinding and failing thereby may
3854 suspend temporarily, and may involve (recursive) unwinding of the stack of a
3855 failed destructor. Nonetheless, the outermost unwinding activity will continue
3856 until the stack is unwound and the task transitions to the *dead*
3857 state. There is no way to "recover" from task failure. Once a task has
3858 temporarily suspended its unwinding in the *failing* state, failure
3859 occurring from within this destructor results in *hard* failure.
3860 A hard failure currently results in the process aborting.
3862 A task in the *dead* state cannot transition to other states; it exists
3863 only to have its termination status inspected by other tasks, and/or to await
3864 reclamation when the last reference to it drops.
3868 The currently scheduled task is given a finite *time slice* in which to
3869 execute, after which it is *descheduled* at a loop-edge or similar
3870 preemption point, and another task within is scheduled, pseudo-randomly.
3872 An executing task can yield control at any time, by making a library call to
3873 `std::task::yield`, which deschedules it immediately. Entering any other
3874 non-executing state (blocked, dead) similarly deschedules the task.
3876 # Runtime services, linkage and debugging
3878 The Rust _runtime_ is a relatively compact collection of C++ and Rust code
3879 that provides fundamental services and datatypes to all Rust tasks at
3880 run-time. It is smaller and simpler than many modern language runtimes. It is
3881 tightly integrated into the language's execution model of memory, tasks,
3882 communication and logging.
3884 > **Note:** The runtime library will merge with the `std` library in future versions of Rust.
3886 ### Memory allocation
3888 The runtime memory-management system is based on a _service-provider interface_,
3889 through which the runtime requests blocks of memory from its environment
3890 and releases them back to its environment when they are no longer needed.
3891 The default implementation of the service-provider interface
3892 consists of the C runtime functions `malloc` and `free`.
3894 The runtime memory-management system, in turn, supplies Rust tasks with
3895 facilities for allocating releasing stacks, as well as allocating and freeing
3900 The runtime provides C and Rust code to assist with various built-in types,
3901 such as vectors, strings, and the low level communication system (ports,
3904 Support for other built-in types such as simple types, tuples, records, and
3905 enums is open-coded by the Rust compiler.
3907 ### Task scheduling and communication
3909 The runtime provides code to manage inter-task communication. This includes
3910 the system of task-lifecycle state transitions depending on the contents of
3911 queues, as well as code to copy values between queues and their recipients and
3912 to serialize values for transmission over operating-system inter-process
3913 communication facilities.
3917 The Rust compiler supports various methods to link crates together both
3918 statically and dynamically. This section will explore the various methods to
3919 link Rust crates together, and more information about native libraries can be
3920 found in the [ffi tutorial][ffi].
3922 In one session of compilation, the compiler can generate multiple artifacts
3923 through the usage of either command line flags or the `crate_type` attribute.
3924 If one or more command line flag is specified, all `crate_type` attributes will
3925 be ignored in favor of only building the artifacts specified by command line.
3927 * `--crate-type=bin`, `#[crate_type = "bin"]` - A runnable executable will be
3928 produced. This requires that there is a `main` function in the crate which
3929 will be run when the program begins executing. This will link in all Rust and
3930 native dependencies, producing a distributable binary.
3932 * `--crate-type=lib`, `#[crate_type = "lib"]` - A Rust library will be produced.
3933 This is an ambiguous concept as to what exactly is produced because a library
3934 can manifest itself in several forms. The purpose of this generic `lib` option
3935 is to generate the "compiler recommended" style of library. The output library
3936 will always be usable by rustc, but the actual type of library may change from
3937 time-to-time. The remaining output types are all different flavors of
3938 libraries, and the `lib` type can be seen as an alias for one of them (but the
3939 actual one is compiler-defined).
3941 * `--crate-type=dylib`, `#[crate_type = "dylib"]` - A dynamic Rust library will
3942 be produced. This is different from the `lib` output type in that this forces
3943 dynamic library generation. The resulting dynamic library can be used as a
3944 dependency for other libraries and/or executables. This output type will
3945 create `*.so` files on linux, `*.dylib` files on osx, and `*.dll` files on
3948 * `--crate-type=staticlib`, `#[crate_type = "staticlib"]` - A static system
3949 library will be produced. This is different from other library outputs in that
3950 the Rust compiler will never attempt to link to `staticlib` outputs. The
3951 purpose of this output type is to create a static library containing all of
3952 the local crate's code along with all upstream dependencies. The static
3953 library is actually a `*.a` archive on linux and osx and a `*.lib` file on
3954 windows. This format is recommended for use in situtations such as linking
3955 Rust code into an existing non-Rust application because it will not have
3956 dynamic dependencies on other Rust code.
3958 * `--crate-type=rlib`, `#[crate_type = "rlib"]` - A "Rust library" file will be
3959 produced. This is used as an intermediate artifact and can be thought of as a
3960 "static Rust library". These `rlib` files, unlike `staticlib` files, are
3961 interpreted by the Rust compiler in future linkage. This essentially means
3962 that `rustc` will look for metadata in `rlib` files like it looks for metadata
3963 in dynamic libraries. This form of output is used to produce statically linked
3964 executables as well as `staticlib` outputs.
3966 Note that these outputs are stackable in the sense that if multiple are
3967 specified, then the compiler will produce each form of output at once without
3968 having to recompile. However, this only applies for outputs specified by the same
3969 method. If only `crate_type` attributes are specified, then they will all be
3970 built, but if one or more `--crate-type` command line flag is specified,
3971 then only those outputs will be built.
3973 With all these different kinds of outputs, if crate A depends on crate B, then
3974 the compiler could find B in various different forms throughout the system. The
3975 only forms looked for by the compiler, however, are the `rlib` format and the
3976 dynamic library format. With these two options for a dependent library, the
3977 compiler must at some point make a choice between these two formats. With this
3978 in mind, the compiler follows these rules when determining what format of
3979 dependencies will be used:
3981 1. If a dynamic library is being produced, then it is required for all upstream
3982 Rust dependencies to also be dynamic. This is a limitation of the current
3983 implementation of the linkage model. The reason behind this limitation is to
3984 prevent multiple copies of the same upstream library from showing up, and in
3985 the future it is planned to support a mixture of dynamic and static linking.
3987 When producing a dynamic library, the compiler will generate an error if an
3988 upstream dependency could not be found, and also if an upstream dependency
3989 could only be found in an `rlib` format. Remember that `staticlib` formats
3990 are always ignored by `rustc` for crate-linking purposes.
3992 2. If a static library is being produced, all upstream dependecies are
3993 required to be available in `rlib` formats. This requirement stems from the
3994 same reasons that a dynamic library must have all dynamic dependencies.
3996 Note that it is impossible to link in native dynamic dependencies to a static
3997 library, and in this case warnings will be printed about all unlinked native
3998 dynamic dependencies.
4000 3. If an `rlib` file is being produced, then there are no restrictions on what
4001 format the upstream dependencies are available in. It is simply required that
4002 all upstream dependencies be available for reading metadata from.
4004 The reason for this is that `rlib` files do not contain any of their upstream
4005 dependencies. It wouldn't be very efficient for all `rlib` files to contain a
4006 copy of `libstd.rlib`!
4008 4. If an executable is being produced, then things get a little interesting. As
4009 with the above limitations in dynamic and static libraries, it is required
4010 for all upstream dependencies to be in the same format. The next question is
4011 whether to prefer a dynamic or a static format. The compiler currently favors
4012 static linking over dynamic linking, but this can be inverted with the `-C
4013 prefer-dynamic` flag to the compiler.
4015 What this means is that first the compiler will attempt to find all upstream
4016 dependencies as `rlib` files, and if successful, it will create a statically
4017 linked executable. If an upstream dependency is missing as an `rlib` file,
4018 then the compiler will force all dependencies to be dynamic and will generate
4019 errors if dynamic versions could not be found.
4021 In general, `--crate-type=bin` or `--crate-type=lib` should be sufficient for
4022 all compilation needs, and the other options are just available if more
4023 fine-grained control is desired over the output format of a Rust crate.
4027 The runtime contains a system for directing [logging
4028 expressions](#logging-expressions) to a logging console and/or internal logging
4029 buffers. Logging can be enabled per module.
4031 Logging output is enabled by setting the `RUST_LOG` environment
4032 variable. `RUST_LOG` accepts a logging specification made up of a
4033 comma-separated list of paths, with optional log levels. For each
4034 module containing log expressions, if `RUST_LOG` contains the path to
4035 that module or a parent of that module, then logs of the appropriate
4036 level will be output to the console.
4038 The path to a module consists of the crate name, any parent modules,
4039 then the module itself, all separated by double colons (`::`). The
4040 optional log level can be appended to the module path with an equals
4041 sign (`=`) followed by the log level, from 1 to 4, inclusive. Level 1
4042 is the error level, 2 is warning, 3 info, and 4 debug. You can also
4043 use the symbolic constants `error`, `warn`, `info`, and `debug`. Any
4044 logs less than or equal to the specified level will be output. If not
4045 specified then log level 4 is assumed. Debug messages can be omitted
4046 by passing `--cfg ndebug` to `rustc`.
4048 As an example, to see all the logs generated by the compiler, you would set
4049 `RUST_LOG` to `rustc`, which is the crate name (as specified in its `crate_id`
4050 [attribute](#attributes)). To narrow down the logs to just crate resolution,
4051 you would set it to `rustc::metadata::creader`. To see just error logging
4054 Note that when compiling source files that don't specify a
4055 crate name the crate is given a default name that matches the source file,
4056 with the extension removed. In that case, to turn on logging for a program
4057 compiled from, e.g. `helloworld.rs`, `RUST_LOG` should be set to `helloworld`.
4059 #### Logging Expressions
4061 Rust provides several macros to log information. Here's a simple Rust program
4062 that demonstrates all four of them:
4066 #[phase(syntax, link)] extern crate log;
4069 error!("This is an error log")
4070 warn!("This is a warn log")
4071 info!("this is an info log")
4072 debug!("This is a debug log")
4076 These four log levels correspond to levels 1-4, as controlled by `RUST_LOG`:
4078 ``` {.bash .notrust}
4079 $ RUST_LOG=rust=3 ./rust
4080 This is an error log
4085 # Appendix: Rationales and design tradeoffs
4089 # Appendix: Influences and further references
4093 > The essential problem that must be solved in making a fault-tolerant
4094 > software system is therefore that of fault-isolation. Different programmers
4095 > will write different modules, some modules will be correct, others will have
4096 > errors. We do not want the errors in one module to adversely affect the
4097 > behaviour of a module which does not have any errors.
4099 > — Joe Armstrong
4101 > In our approach, all data is private to some process, and processes can
4102 > only communicate through communications channels. *Security*, as used
4103 > in this paper, is the property which guarantees that processes in a system
4104 > cannot affect each other except by explicit communication.
4106 > When security is absent, nothing which can be proven about a single module
4107 > in isolation can be guaranteed to hold when that module is embedded in a
4110 > — Robert Strom and Shaula Yemini
4112 > Concurrent and applicative programming complement each other. The
4113 > ability to send messages on channels provides I/O without side effects,
4114 > while the avoidance of shared data helps keep concurrent processes from
4119 Rust is not a particularly original language. It may however appear unusual
4120 by contemporary standards, as its design elements are drawn from a number of
4121 "historical" languages that have, with a few exceptions, fallen out of
4122 favour. Five prominent lineages contribute the most, though their influences
4123 have come and gone during the course of Rust's development:
4125 * The NIL (1981) and Hermes (1990) family. These languages were developed by
4126 Robert Strom, Shaula Yemini, David Bacon and others in their group at IBM
4127 Watson Research Center (Yorktown Heights, NY, USA).
4129 * The Erlang (1987) language, developed by Joe Armstrong, Robert Virding, Claes
4130 Wikström, Mike Williams and others in their group at the Ericsson Computer
4131 Science Laboratory (Älvsjö, Stockholm, Sweden) .
4133 * The Sather (1990) language, developed by Stephen Omohundro, Chu-Cheow Lim,
4134 Heinz Schmidt and others in their group at The International Computer
4135 Science Institute of the University of California, Berkeley (Berkeley, CA,
4138 * The Newsqueak (1988), Alef (1995), and Limbo (1996) family. These
4139 languages were developed by Rob Pike, Phil Winterbottom, Sean Dorward and
4140 others in their group at Bell Labs Computing Sciences Research Center
4141 (Murray Hill, NJ, USA).
4143 * The Napier (1985) and Napier88 (1988) family. These languages were
4144 developed by Malcolm Atkinson, Ron Morrison and others in their group at
4145 the University of St. Andrews (St. Andrews, Fife, UK).
4147 Additional specific influences can be seen from the following languages:
4149 * The structural algebraic types and compilation manager of SML.
4150 * The attribute and assembly systems of C#.
4151 * The references and deterministic destructor system of C++.
4152 * The memory region systems of the ML Kit and Cyclone.
4153 * The typeclass system of Haskell.
4154 * The lexical identifier rule of Python.
4155 * The block syntax of Ruby.
4157 [ffi]: guide-ffi.html