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]
21 library 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
30 Rust is a work in progress. The language continues to evolve as the design
31 shifts and is fleshed out in working code. Certain parts work, certain parts
32 do not, certain parts will be removed or changed.
34 This manual is a snapshot written in the present tense. All features described
35 exist in working code unless otherwise noted, but some are quite primitive or
36 remain to be further modified by planned work. Some may be temporary. It is a
37 *draft*, and we ask that you not take anything you read here as final.
39 If you have suggestions to make, please try to focus them on *reductions* to
40 the language: possible features that can be combined or omitted. We aim to
41 keep the size and complexity of the language under control.
43 > **Note:** The grammar for Rust given in this document is rough and
44 > very incomplete; only a modest number of sections have accompanying grammar
45 > rules. Formalizing the grammar accepted by the Rust parser is ongoing work,
46 > but future versions of this document will contain a complete
47 > grammar. Moreover, we hope that this grammar will be extracted and verified
48 > as LL(1) by an automated grammar-analysis tool, and further tested against the
49 > Rust sources. Preliminary versions of this automation exist, but are not yet
54 Rust's grammar is defined over Unicode codepoints, each conventionally
55 denoted `U+XXXX`, for 4 or more hexadecimal digits `X`. _Most_ of Rust's
56 grammar is confined to the ASCII range of Unicode, and is described in this
57 document by a dialect of Extended Backus-Naur Form (EBNF), specifically a
58 dialect of EBNF supported by common automated LL(k) parsing tools such as
59 `llgen`, rather than the dialect given in ISO 14977. The dialect can be
60 defined self-referentially as follows:
62 ~~~~ {.ebnf .notation}
64 rule : nonterminal ':' productionrule ';' ;
65 productionrule : production [ '|' production ] * ;
67 term : element repeats ;
68 element : LITERAL | IDENTIFIER | '[' productionrule ']' ;
69 repeats : [ '*' | '+' ] NUMBER ? | NUMBER ? | '?' ;
74 - Whitespace in the grammar is ignored.
75 - Square brackets are used to group rules.
76 - `LITERAL` is a single printable ASCII character, or an escaped hexadecimal
77 ASCII code of the form `\xQQ`, in single quotes, denoting the corresponding
78 Unicode codepoint `U+00QQ`.
79 - `IDENTIFIER` is a nonempty string of ASCII letters and underscores.
80 - The `repeat` forms apply to the adjacent `element`, and are as follows:
81 - `?` means zero or one repetition
82 - `*` means zero or more repetitions
83 - `+` means one or more repetitions
84 - NUMBER trailing a repeat symbol gives a maximum repetition count
85 - NUMBER on its own gives an exact repetition count
87 This EBNF dialect should hopefully be familiar to many readers.
89 ## Unicode productions
91 A few productions in Rust's grammar permit Unicode codepoints outside the ASCII range.
92 We define these productions in terms of character properties specified in the Unicode standard,
93 rather than in terms of ASCII-range codepoints.
94 The section [Special Unicode Productions](#special-unicode-productions) lists these productions.
96 ## String table productions
98 Some rules in the grammar — notably [unary
99 operators](#unary-operator-expressions), [binary
100 operators](#binary-operator-expressions), and [keywords](#keywords) —
101 are given in a simplified form: as a listing of a table of unquoted,
102 printable whitespace-separated strings. These cases form a subset of
103 the rules regarding the [token](#tokens) rule, and are assumed to be
104 the result of a lexical-analysis phase feeding the parser, driven by a
105 DFA, operating over the disjunction of all such string table entries.
107 When such a string enclosed in double-quotes (`"`) occurs inside the
108 grammar, it is an implicit reference to a single member of such a string table
109 production. See [tokens](#tokens) for more information.
115 Rust input is interpreted as a sequence of Unicode codepoints encoded in UTF-8,
116 normalized to Unicode normalization form NFKC.
117 Most Rust grammar rules are defined in terms of printable ASCII-range codepoints,
118 but a small number are defined in terms of Unicode properties or explicit
119 codepoint lists. [^inputformat]
121 [^inputformat]: Substitute definitions for the special Unicode productions are
122 provided to the grammar verifier, restricted to ASCII range, when verifying
123 the grammar in this document.
125 ## Special Unicode Productions
127 The following productions in the Rust grammar are defined in terms of Unicode properties:
128 `ident`, `non_null`, `non_star`, `non_eol`, `non_slash_or_star`, `non_single_quote` and `non_double_quote`.
132 The `ident` production is any nonempty Unicode string of the following form:
134 - The first character has property `XID_start`
135 - The remaining characters have property `XID_continue`
137 that does _not_ occur in the set of [keywords](#keywords).
139 Note: `XID_start` and `XID_continue` as character properties cover the
140 character ranges used to form the more familiar C and Java language-family
143 ### Delimiter-restricted productions
145 Some productions are defined by exclusion of particular Unicode characters:
147 - `non_null` is any single Unicode character aside from `U+0000` (null)
148 - `non_eol` is `non_null` restricted to exclude `U+000A` (`'\n'`)
149 - `non_star` is `non_null` restricted to exclude `U+002A` (`*`)
150 - `non_slash_or_star` is `non_null` restricted to exclude `U+002F` (`/`) and `U+002A` (`*`)
151 - `non_single_quote` is `non_null` restricted to exclude `U+0027` (`'`)
152 - `non_double_quote` is `non_null` restricted to exclude `U+0022` (`"`)
157 comment : block_comment | line_comment ;
158 block_comment : "/*" block_comment_body * '*' + '/' ;
159 block_comment_body : [block_comment | character] * ;
160 line_comment : "//" non_eol * ;
163 Comments in Rust code follow the general C++ style of line and block-comment forms.
164 Nested block comments are supported.
166 Line comments beginning with exactly _three_ slashes (`///`), and block
167 comments beginning with exactly one repeated asterisk in the block-open
168 sequence (`/**`), are interpreted as a special syntax for `doc`
169 [attributes](#attributes). That is, they are equivalent to writing
170 `#[doc="..."]` around the body of the comment (this includes the comment
171 characters themselves, ie `/// Foo` turns into `#[doc="/// Foo"]`).
173 Non-doc comments are interpreted as a form of whitespace.
178 whitespace_char : '\x20' | '\x09' | '\x0a' | '\x0d' ;
179 whitespace : [ whitespace_char | comment ] + ;
182 The `whitespace_char` production is any nonempty Unicode string consisting of any
183 of the following Unicode characters: `U+0020` (space, `' '`), `U+0009` (tab,
184 `'\t'`), `U+000A` (LF, `'\n'`), `U+000D` (CR, `'\r'`).
186 Rust is a "free-form" language, meaning that all forms of whitespace serve
187 only to separate _tokens_ in the grammar, and have no semantic significance.
189 A Rust program has identical meaning if each whitespace element is replaced
190 with any other legal whitespace element, such as a single space character.
195 simple_token : keyword | unop | binop ;
196 token : simple_token | ident | literal | symbol | whitespace token ;
199 Tokens are primitive productions in the grammar defined by regular
200 (non-recursive) languages. "Simple" tokens are given in [string table
201 production](#string-table-productions) form, and occur in the rest of the
202 grammar as double-quoted strings. Other tokens have exact rules given.
206 The keywords are the following strings:
208 ~~~~ {.text .keyword}
219 self static struct super
225 Each of these keywords has special meaning in its grammar,
226 and all of them are excluded from the `ident` rule.
230 A literal is an expression consisting of a single token, rather than a
231 sequence of tokens, that immediately and directly denotes the value it
232 evaluates to, rather than referring to it by name or some other evaluation
233 rule. A literal is a form of constant expression, so is evaluated (primarily)
237 literal : string_lit | char_lit | byte_string_lit | byte_lit | num_lit ;
240 #### Character and string literals
243 char_lit : '\x27' char_body '\x27' ;
244 string_lit : '"' string_body * '"' | 'r' raw_string ;
246 char_body : non_single_quote
247 | '\x5c' [ '\x27' | common_escape | unicode_escape ] ;
249 string_body : non_double_quote
250 | '\x5c' [ '\x22' | common_escape | unicode_escape ] ;
251 raw_string : '"' raw_string_body '"' | '#' raw_string '#' ;
253 common_escape : '\x5c'
254 | 'n' | 'r' | 't' | '0'
256 unicode_escape : 'u' hex_digit 4
259 hex_digit : 'a' | 'b' | 'c' | 'd' | 'e' | 'f'
260 | 'A' | 'B' | 'C' | 'D' | 'E' | 'F'
262 oct_digit : '0' | '1' | '2' | '3' | '4' | '5' | '6' | '7' ;
263 dec_digit : '0' | nonzero_dec ;
264 nonzero_dec: '1' | '2' | '3' | '4'
265 | '5' | '6' | '7' | '8' | '9' ;
268 A _character literal_ is a single Unicode character enclosed within two
269 `U+0027` (single-quote) characters, with the exception of `U+0027` itself,
270 which must be _escaped_ by a preceding U+005C character (`\`).
272 A _string literal_ is a sequence of any Unicode characters enclosed within
273 two `U+0022` (double-quote) characters, with the exception of `U+0022`
274 itself, which must be _escaped_ by a preceding `U+005C` character (`\`),
275 or a _raw string literal_.
277 Some additional _escapes_ are available in either character or non-raw string
278 literals. An escape starts with a `U+005C` (`\`) and continues with one of
281 * An _8-bit codepoint escape_ escape starts with `U+0078` (`x`) and is
282 followed by exactly two _hex digits_. It denotes the Unicode codepoint
283 equal to the provided hex value.
284 * A _16-bit codepoint escape_ starts with `U+0075` (`u`) and is followed
285 by exactly four _hex digits_. It denotes the Unicode codepoint equal to
286 the provided hex value.
287 * A _32-bit codepoint escape_ starts with `U+0055` (`U`) and is followed
288 by exactly eight _hex digits_. It denotes the Unicode codepoint equal to
289 the provided hex value.
290 * A _whitespace escape_ is one of the characters `U+006E` (`n`), `U+0072`
291 (`r`), or `U+0074` (`t`), denoting the unicode values `U+000A` (LF),
292 `U+000D` (CR) or `U+0009` (HT) respectively.
293 * The _backslash escape_ is the character `U+005C` (`\`) which must be
294 escaped in order to denote *itself*.
296 Raw string literals do not process any escapes. They start with the character
297 `U+0072` (`r`), followed by zero or more of the character `U+0023` (`#`) and a
298 `U+0022` (double-quote) character. The _raw string body_ is not defined in the
299 EBNF grammar above: it can contain any sequence of Unicode characters and is
300 terminated only by another `U+0022` (double-quote) character, followed by the
301 same number of `U+0023` (`#`) characters that preceded the opening `U+0022`
302 (double-quote) character.
304 All Unicode characters contained in the raw string body represent themselves,
305 the characters `U+0022` (double-quote) (except when followed by at least as
306 many `U+0023` (`#`) characters as were used to start the raw string literal) or
307 `U+005C` (`\`) do not have any special meaning.
309 Examples for string literals:
312 "foo"; r"foo"; // foo
313 "\"foo\""; r#""foo""#; // "foo"
316 r##"foo #"# bar"##; // foo #"# bar
318 "\x52"; "R"; r"R"; // R
319 "\\x52"; r"\x52"; // \x52
322 #### Byte and byte string literals
325 byte_lit : 'b' '\x27' byte_body '\x27' ;
326 byte_string_lit : 'b' '"' string_body * '"' | 'b' 'r' raw_byte_string ;
328 byte_body : ascii_non_single_quote
329 | '\x5c' [ '\x27' | common_escape ] ;
331 byte_string_body : ascii_non_double_quote
332 | '\x5c' [ '\x22' | common_escape ] ;
333 raw_byte_string : '"' raw_byte_string_body '"' | '#' raw_byte_string '#' ;
337 A _byte literal_ is a single ASCII character (in the `U+0000` to `U+007F` range)
338 enclosed within two `U+0027` (single-quote) characters,
339 with the exception of `U+0027` itself,
340 which must be _escaped_ by a preceding U+005C character (`\`),
341 or a single _escape_.
342 It is equivalent to a `u8` unsigned 8-bit integer _number literal_.
344 A _byte string literal_ is a sequence of ASCII characters and _escapes_
345 enclosed within two `U+0022` (double-quote) characters,
346 with the exception of `U+0022` itself,
347 which must be _escaped_ by a preceding `U+005C` character (`\`),
348 or a _raw byte string literal_.
349 It is equivalent to a `&'static [u8]` borrowed vectior unsigned 8-bit integers.
351 Some additional _escapes_ are available in either byte or non-raw byte string
352 literals. An escape starts with a `U+005C` (`\`) and continues with one of
355 * An _byte escape_ escape starts with `U+0078` (`x`) and is
356 followed by exactly two _hex digits_. It denotes the byte
357 equal to the provided hex value.
358 * A _whitespace escape_ is one of the characters `U+006E` (`n`), `U+0072`
359 (`r`), or `U+0074` (`t`), denoting the bytes values `0x0A` (ASCII LF),
360 `0x0D` (ASCII CR) or `0x09` (ASCII HT) respectively.
361 * The _backslash escape_ is the character `U+005C` (`\`) which must be
362 escaped in order to denote its ASCII encoding `0x5C`.
364 Raw byte string literals do not process any escapes.
365 They start with the character `U+0072` (`r`),
366 followed by `U+0062` (`b`),
367 followed by zero or more of the character `U+0023` (`#`),
368 and a `U+0022` (double-quote) character.
369 The _raw string body_ is not defined in the EBNF grammar above:
370 it can contain any sequence of ASCII characters and is
371 terminated only by another `U+0022` (double-quote) character, followed by the
372 same number of `U+0023` (`#`) characters that preceded the opening `U+0022`
373 (double-quote) character.
374 A raw byte string literal can not contain any non-ASCII byte.
376 All characters contained in the raw string body represent their ASCII encoding,
377 the characters `U+0022` (double-quote) (except when followed by at least as
378 many `U+0023` (`#`) characters as were used to start the raw string literal) or
379 `U+005C` (`\`) do not have any special meaning.
381 Examples for byte string literals:
384 b"foo"; br"foo"; // foo
385 b"\"foo\""; br#""foo""#; // "foo"
388 br##"foo #"# bar"##; // foo #"# bar
390 b"\x52"; b"R"; br"R"; // R
391 b"\\x52"; br"\x52"; // \x52
397 num_lit : nonzero_dec [ dec_digit | '_' ] * num_suffix ?
398 | '0' [ [ dec_digit | '_' ] * num_suffix ?
399 | 'b' [ '1' | '0' | '_' ] + int_suffix ?
400 | 'o' [ oct_digit | '_' ] + int_suffix ?
401 | 'x' [ hex_digit | '_' ] + int_suffix ? ] ;
403 num_suffix : int_suffix | float_suffix ;
405 int_suffix : 'u' int_suffix_size ?
406 | 'i' int_suffix_size ? ;
407 int_suffix_size : [ '8' | '1' '6' | '3' '2' | '6' '4' ] ;
409 float_suffix : [ exponent | '.' dec_lit exponent ? ] ? float_suffix_ty ? ;
410 float_suffix_ty : 'f' [ '3' '2' | '6' '4' ] ;
411 exponent : ['E' | 'e'] ['-' | '+' ] ? dec_lit ;
412 dec_lit : [ dec_digit | '_' ] + ;
415 A _number literal_ is either an _integer literal_ or a _floating-point
416 literal_. The grammar for recognizing the two kinds of literals is mixed,
417 as they are differentiated by suffixes.
419 ##### Integer literals
421 An _integer literal_ has one of four forms:
423 * A _decimal literal_ starts with a *decimal digit* and continues with any
424 mixture of *decimal digits* and _underscores_.
425 * A _hex literal_ starts with the character sequence `U+0030` `U+0078`
426 (`0x`) and continues as any mixture hex digits and underscores.
427 * An _octal literal_ starts with the character sequence `U+0030` `U+006F`
428 (`0o`) and continues as any mixture octal digits and underscores.
429 * A _binary literal_ starts with the character sequence `U+0030` `U+0062`
430 (`0b`) and continues as any mixture binary digits and underscores.
432 An integer literal may be followed (immediately, without any spaces) by an
433 _integer suffix_, which changes the type of the literal. There are two kinds
434 of integer literal suffix:
436 * The `i` and `u` suffixes give the literal type `int` or `uint`,
438 * Each of the signed and unsigned machine types `u8`, `i8`,
439 `u16`, `i16`, `u32`, `i32`, `u64` and `i64`
440 give the literal the corresponding machine type.
442 The type of an _unsuffixed_ integer literal is determined by type inference.
443 If an integer type can be _uniquely_ determined from the surrounding program
444 context, the unsuffixed integer literal has that type. If the program context
445 underconstrains the type, the unsuffixed integer literal's type is `int`; if
446 the program context overconstrains the type, it is considered a static type
449 Examples of integer literals of various forms:
452 123; 0xff00; // type determined by program context
453 // defaults to int in absence of type
459 0o70_i16; // type i16
460 0b1111_1111_1001_0000_i32; // type i32
463 ##### Floating-point literals
465 A _floating-point literal_ has one of two forms:
467 * Two _decimal literals_ separated by a period
468 character `U+002E` (`.`), with an optional _exponent_ trailing after the
469 second decimal literal.
470 * A single _decimal literal_ followed by an _exponent_.
472 By default, a floating-point literal has a generic type, but will fall back to
473 `f64`. A floating-point literal may be followed (immediately, without any
474 spaces) by a _floating-point suffix_, which changes the type of the literal.
475 There are two floating-point suffixes: `f32`, and `f64` (the 32-bit and 64-bit
476 floating point types).
478 Examples of floating-point literals of various forms:
484 12E+99_f64; // type f64
487 ##### Unit and boolean literals
489 The _unit value_, the only value of the type that has the same name, is written as `()`.
490 The two values of the boolean type are written `true` and `false`.
496 | '#' | '[' | ']' | '(' | ')' | '{' | '}'
500 Symbols are a general class of printable [token](#tokens) that play structural
501 roles in a variety of grammar productions. They are catalogued here for
502 completeness as the set of remaining miscellaneous printable tokens that do not
503 otherwise appear as [unary operators](#unary-operator-expressions), [binary
504 operators](#binary-operator-expressions), or [keywords](#keywords).
510 expr_path : [ "::" ] ident [ "::" expr_path_tail ] + ;
511 expr_path_tail : '<' type_expr [ ',' type_expr ] + '>'
514 type_path : ident [ type_path_tail ] + ;
515 type_path_tail : '<' type_expr [ ',' type_expr ] + '>'
519 A _path_ is a sequence of one or more path components _logically_ separated by
520 a namespace qualifier (`::`). If a path consists of only one component, it may
521 refer to either an [item](#items) or a [slot](#memory-slots) in a local
522 control scope. If a path has multiple components, it refers to an item.
524 Every item has a _canonical path_ within its crate, but the path naming an
525 item is only meaningful within a given crate. There is no global namespace
526 across crates; an item's canonical path merely identifies it within the crate.
528 Two examples of simple paths consisting of only identifier components:
535 Path components are usually [identifiers](#identifiers), but the trailing
536 component of a path may be an angle-bracket-enclosed list of type
537 arguments. In [expression](#expressions) context, the type argument list is
538 given after a final (`::`) namespace qualifier in order to disambiguate it
539 from a relational expression involving the less-than symbol (`<`). In type
540 expression context, the final namespace qualifier is omitted.
542 Two examples of paths with type arguments:
545 # struct HashMap<K, V>;
547 # fn id<T>(t: T) -> T { t }
548 type T = HashMap<int,String>; // Type arguments used in a type expression
549 let x = id::<int>(10); // Type arguments used in a call expression
553 Paths can be denoted with various leading qualifiers to change the meaning of
556 * Paths starting with `::` are considered to be global paths where the
557 components of the path start being resolved from the crate root. Each
558 identifier in the path must resolve to an item.
566 ::a::foo(); // call a's foo function
572 * Paths starting with the keyword `super` begin resolution relative to the
573 parent module. Each further identifier must resolve to an item
581 super::a::foo(); // call a's foo function
587 * Paths starting with the keyword `self` begin resolution relative to the
588 current module. Each further identifier must resolve to an item.
600 A number of minor features of Rust are not central enough to have their own
601 syntax, and yet are not implementable as functions. Instead, they are given
602 names, and invoked through a consistent syntax: `name!(...)`. Examples
605 * `format!` : format data into a string
606 * `env!` : look up an environment variable's value at compile time
607 * `file!`: return the path to the file being compiled
608 * `stringify!` : pretty-print the Rust expression given as an argument
609 * `include!` : include the Rust expression in the given file
610 * `include_str!` : include the contents of the given file as a string
611 * `include_bin!` : include the contents of the given file as a binary blob
612 * `error!`, `warn!`, `info!`, `debug!` : provide diagnostic information.
614 All of the above extensions are expressions with values.
619 expr_macro_rules : "macro_rules" '!' ident '(' macro_rule * ')' ;
620 macro_rule : '(' matcher * ')' "=>" '(' transcriber * ')' ';' ;
621 matcher : '(' matcher * ')' | '[' matcher * ']'
622 | '{' matcher * '}' | '$' ident ':' ident
623 | '$' '(' matcher * ')' sep_token? [ '*' | '+' ]
624 | non_special_token ;
625 transcriber : '(' transcriber * ')' | '[' transcriber * ']'
626 | '{' transcriber * '}' | '$' ident
627 | '$' '(' transcriber * ')' sep_token? [ '*' | '+' ]
628 | non_special_token ;
631 User-defined syntax extensions are called "macros",
632 and the `macro_rules` syntax extension defines them.
633 Currently, user-defined macros can expand to expressions, statements, or items.
635 (A `sep_token` is any token other than `*` and `+`.
636 A `non_special_token` is any token other than a delimiter or `$`.)
638 The macro expander looks up macro invocations by name,
639 and tries each macro rule in turn.
640 It transcribes the first successful match.
641 Matching and transcription are closely related to each other,
642 and we will describe them together.
646 The macro expander matches and transcribes every token that does not begin with a `$` literally, including delimiters.
647 For parsing reasons, delimiters must be balanced, but they are otherwise not special.
649 In the matcher, `$` _name_ `:` _designator_ matches the nonterminal in the
650 Rust syntax named by _designator_. Valid designators are `item`, `block`,
651 `stmt`, `pat`, `expr`, `ty` (type), `ident`, `path`, `matchers` (lhs of the `=>` in macro rules),
652 `tt` (rhs of the `=>` in macro rules). In the transcriber, the designator is already known, and so only
653 the name of a matched nonterminal comes after the dollar sign.
655 In both the matcher and transcriber, the Kleene star-like operator indicates repetition.
656 The Kleene star operator consists of `$` and parens, optionally followed by a separator token, followed by `*` or `+`.
657 `*` means zero or more repetitions, `+` means at least one repetition.
658 The parens are not matched or transcribed.
659 On the matcher side, a name is bound to _all_ of the names it
660 matches, in a structure that mimics the structure of the repetition
661 encountered on a successful match. The job of the transcriber is to sort that
664 The rules for transcription of these repetitions are called "Macro By Example".
665 Essentially, one "layer" of repetition is discharged at a time, and all of
666 them must be discharged by the time a name is transcribed. Therefore,
667 `( $( $i:ident ),* ) => ( $i )` is an invalid macro, but
668 `( $( $i:ident ),* ) => ( $( $i:ident ),* )` is acceptable (if trivial).
670 When Macro By Example encounters a repetition, it examines all of the `$`
671 _name_ s that occur in its body. At the "current layer", they all must repeat
672 the same number of times, so
673 ` ( $( $i:ident ),* ; $( $j:ident ),* ) => ( $( ($i,$j) ),* )` is valid if
674 given the argument `(a,b,c ; d,e,f)`, but not `(a,b,c ; d,e)`. The repetition
675 walks through the choices at that layer in lockstep, so the former input
676 transcribes to `( (a,d), (b,e), (c,f) )`.
678 Nested repetitions are allowed.
680 ### Parsing limitations
682 The parser used by the macro system is reasonably powerful, but the parsing of
683 Rust syntax is restricted in two ways:
685 1. The parser will always parse as much as possible. If it attempts to match
686 `$i:expr [ , ]` against `8 [ , ]`, it will attempt to parse `i` as an array
687 index operation and fail. Adding a separator can solve this problem.
688 2. The parser must have eliminated all ambiguity by the time it reaches a `$` _name_ `:` _designator_.
689 This requirement most often affects name-designator pairs when they occur at the beginning of, or immediately after, a `$(...)*`; requiring a distinctive token in front can solve the problem.
691 ## Syntax extensions useful for the macro author
693 * `log_syntax!` : print out the arguments at compile time
694 * `trace_macros!` : supply `true` or `false` to enable or disable macro expansion logging
695 * `stringify!` : turn the identifier argument into a string literal
696 * `concat!` : concatenates a comma-separated list of literals
697 * `concat_idents!` : create a new identifier by concatenating the arguments
699 # Crates and source files
701 Rust is a *compiled* language.
702 Its semantics obey a *phase distinction* between compile-time and run-time.
703 Those semantic rules that have a *static interpretation* govern the success or failure of compilation.
704 We refer to these rules as "static semantics".
705 Semantic rules called "dynamic semantics" govern the behavior of programs at run-time.
706 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.
708 The compilation model centres on artifacts called _crates_.
709 Each compilation processes a single crate in source form, and if successful,
710 produces a single crate in binary form: either an executable or a
711 library.[^cratesourcefile]
713 [^cratesourcefile]: A crate is somewhat analogous to an *assembly* in the
714 ECMA-335 CLI model, a *library* in the SML/NJ Compilation Manager, a *unit*
715 in the Owens and Flatt module system, or a *configuration* in Mesa.
717 A _crate_ is a unit of compilation and linking, as well as versioning, distribution and runtime loading.
718 A crate contains a _tree_ of nested [module](#modules) scopes.
719 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.
721 The Rust compiler is always invoked with a single source file as input, and always produces a single output crate.
722 The processing of that source file may result in other source files being loaded as modules.
723 Source files have the extension `.rs`.
725 A Rust source file describes a module, the name and
726 location of which — in the module tree of the current crate — are defined
727 from outside the source file: either by an explicit `mod_item` in
728 a referencing source file, or by the name of the crate itself.
730 Each source file contains a sequence of zero or more `item` definitions,
731 and may optionally begin with any number of `attributes` that apply to the containing module.
732 Attributes on the anonymous crate module define important metadata that influences
733 the behavior of the compiler.
736 # #![allow(unused_attribute)]
738 #![crate_id = "projx#2.5"]
740 // Additional metadata attributes
741 #![desc = "Project X"]
743 #![comment = "This is a comment on Project X."]
745 // Specify the output type
746 #![crate_type = "lib"]
749 #![warn(non_camel_case_types)]
752 A crate that contains a `main` function can be compiled to an executable.
753 If a `main` function is present, its return type must be [`unit`](#primitive-types) and it must take no arguments.
755 # Items and attributes
757 Crates contain [items](#items),
758 each of which may have some number of [attributes](#attributes) attached to it.
763 item : mod_item | fn_item | type_item | struct_item | enum_item
764 | static_item | trait_item | impl_item | extern_block ;
767 An _item_ is a component of a crate; some module items can be defined in crate
768 files, but most are defined in source files. Items are organized within a
769 crate by a nested set of [modules](#modules). Every crate has a single
770 "outermost" anonymous module; all further items within the crate have
771 [paths](#paths) within the module tree of the crate.
773 Items are entirely determined at compile-time, generally remain fixed during
774 execution, and may reside in read-only memory.
776 There are several kinds of item:
778 * [modules](#modules)
779 * [functions](#functions)
780 * [type definitions](#type-definitions)
781 * [structures](#structures)
782 * [enumerations](#enumerations)
783 * [static items](#static-items)
785 * [implementations](#implementations)
787 Some items form an implicit scope for the declaration of sub-items. In other
788 words, within a function or module, declarations of items can (in many cases)
789 be mixed with the statements, control blocks, and similar artifacts that
790 otherwise compose the item body. The meaning of these scoped items is the same
791 as if the item was declared outside the scope — it is still a static item —
792 except that the item's *path name* within the module namespace is qualified by
793 the name of the enclosing item, or is private to the enclosing item (in the
795 The grammar specifies the exact locations in which sub-item declarations may appear.
799 All items except modules may be *parameterized* by type. Type parameters are
800 given as a comma-separated list of identifiers enclosed in angle brackets
801 (`<...>`), after the name of the item and before its definition.
802 The type parameters of an item are considered "part of the name", not part of the type of the item.
803 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.
804 In practice, the type-inference system can usually infer such argument types from context.
805 There are no general type-parametric types, only type-parametric items.
806 That is, Rust has no notion of type abstraction: there are no first-class "forall" types.
811 mod_item : "mod" ident ( ';' | '{' mod '}' );
812 mod : [ view_item | item ] * ;
815 A module is a container for zero or more [view items](#view-items) and zero or
816 more [items](#items). The view items manage the visibility of the items
817 defined within the module, as well as the visibility of names from outside the
818 module when referenced from inside the module.
820 A _module item_ is a module, surrounded in braces, named, and prefixed with
821 the keyword `mod`. A module item introduces a new, named module into the tree
822 of modules making up a crate. Modules can nest arbitrarily.
824 An example of a module:
828 type Complex = (f64, f64);
829 fn sin(f: f64) -> f64 {
833 fn cos(f: f64) -> f64 {
837 fn tan(f: f64) -> f64 {
844 Modules and types share the same namespace.
845 Declaring a named type that has the same name as a module in scope is forbidden:
846 that is, a type definition, trait, struct, enumeration, or type parameter
847 can't shadow the name of a module in scope, or vice versa.
849 A module without a body is loaded from an external file, by default with the same
850 name as the module, plus the `.rs` extension.
851 When a nested submodule is loaded from an external file,
852 it is loaded from a subdirectory path that mirrors the module hierarchy.
855 // Load the `vec` module from `vec.rs`
859 // Load the `local_data` module from `task/local_data.rs`
864 The directories and files used for loading external file modules can be influenced
865 with the `path` attribute.
868 #[path = "task_files"]
870 // Load the `local_data` module from `task_files/tls.rs`
879 view_item : extern_crate_decl | use_decl ;
882 A view item manages the namespace of a module.
883 View items do not define new items, but rather, simply change other items' visibility.
884 There are several kinds of view item:
886 * [`extern crate` declarations](#extern-crate-declarations)
887 * [`use` declarations](#use-declarations)
889 ##### Extern crate declarations
892 extern_crate_decl : "extern" "crate" ident [ '(' link_attrs ')' ] ? [ '=' string_lit ] ? ;
893 link_attrs : link_attr [ ',' link_attrs ] + ;
894 link_attr : ident '=' literal ;
897 An _`extern crate` declaration_ specifies a dependency on an external crate.
898 The external crate is then bound into the declaring scope as the `ident` provided
899 in the `extern_crate_decl`.
901 The external crate is resolved to a specific `soname` at compile time, and a
902 runtime linkage requirement to that `soname` is passed to the linker for
903 loading at runtime. The `soname` is resolved at compile time by scanning the
904 compiler's library path and matching the optional `crateid` provided as a string literal
905 against the `crateid` attributes that were declared on the external crate when
906 it was compiled. If no `crateid` is provided, a default `name` attribute is
907 assumed, equal to the `ident` given in the `extern_crate_decl`.
909 Four examples of `extern crate` declarations:
914 extern crate std; // equivalent to: extern crate std = "std";
916 extern crate ruststd = "std"; // linking to 'std' under another name
918 extern crate foo = "some/where/rust-foo#foo:1.0"; // a full crate ID for external tools
921 ##### Use declarations
924 use_decl : "pub" ? "use" [ ident '=' path
927 path_glob : ident [ "::" [ path_glob
929 | '{' ident [ ',' ident ] * '}' ;
932 A _use declaration_ creates one or more local name bindings synonymous
933 with some other [path](#paths).
934 Usually a `use` declaration is used to shorten the path required to refer to a
935 module item. These declarations may appear at the top of [modules](#modules) and
938 *Note*: Unlike in many languages,
939 `use` declarations in Rust do *not* declare linkage dependency with external crates.
940 Rather, [`extern crate` declarations](#extern-crate-declarations) declare linkage dependencies.
942 Use declarations support a number of convenient shortcuts:
944 * Rebinding the target name as a new local name, using the syntax `use x = p::q::r;`.
945 * Simultaneously binding a list of paths differing only in their final element,
946 using the glob-like brace syntax `use a::b::{c,d,e,f};`
947 * Binding all paths matching a given prefix, using the asterisk wildcard syntax `use a::b::*;`
949 An example of `use` declarations:
952 use std::iter::range_step;
953 use std::option::{Some, None};
958 // Equivalent to 'std::iter::range_step(0u, 10u, 2u);'
959 range_step(0u, 10u, 2u);
961 // Equivalent to 'foo(vec![std::option::Some(1.0f64),
962 // std::option::None]);'
963 foo(vec![Some(1.0f64), None]);
967 Like items, `use` declarations are private to the containing module, by default.
968 Also like items, a `use` declaration can be public, if qualified by the `pub` keyword.
969 Such a `use` declaration serves to _re-export_ a name.
970 A public `use` declaration can therefore _redirect_ some public name to a different target definition:
971 even a definition with a private canonical path, inside a different module.
972 If a sequence of such redirections form a cycle or cannot be resolved unambiguously,
973 they represent a compile-time error.
975 An example of re-exporting:
980 pub use quux::foo::{bar, baz};
989 In this example, the module `quux` re-exports two public names defined in `foo`.
991 Also note that the paths contained in `use` items are relative to the crate root.
992 So, in the previous example, the `use` refers to `quux::foo::{bar, baz}`, and not simply to `foo::{bar, baz}`.
993 This also means that top-level module declarations should be at the crate root if direct usage
994 of the declared modules within `use` items is desired. It is also possible to use `self` and `super`
995 at the beginning of a `use` item to refer to the current and direct parent modules respectively.
996 All rules regarding accessing declared modules in `use` declarations applies to both module declarations
997 and `extern crate` declarations.
999 An example of what will and will not work for `use` items:
1002 # #![allow(unused_imports)]
1003 use foo::native::start; // good: foo is at the root of the crate
1004 use foo::baz::foobaz; // good: foo is at the root of the crate
1007 extern crate native;
1009 use foo::native::start; // good: foo is at crate root
1010 // use native::start; // bad: native is not at the crate root
1011 use self::baz::foobaz; // good: self refers to module 'foo'
1012 use foo::bar::foobar; // good: foo is at crate root
1019 use super::bar::foobar; // good: super refers to module 'foo'
1029 A _function item_ defines a sequence of [statements](#statements) and an optional final [expression](#expressions), along with a name and a set of parameters.
1030 Functions are declared with the keyword `fn`.
1031 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.
1033 A function may also be copied into a first class *value*, in which case the
1034 value has the corresponding [*function type*](#function-types), and can be
1035 used otherwise exactly as a function item (with a minor additional cost of
1036 calling the function indirectly).
1038 Every control path in a function logically ends with a `return` expression or a
1039 diverging expression. If the outermost block of a function has a
1040 value-producing expression in its final-expression position, that expression
1041 is interpreted as an implicit `return` expression applied to the
1044 An example of a function:
1047 fn add(x: int, y: int) -> int {
1052 As with `let` bindings, function arguments are irrefutable patterns,
1053 so any pattern that is valid in a let binding is also valid as an argument.
1056 fn first((value, _): (int, int)) -> int { value }
1060 #### Generic functions
1062 A _generic function_ allows one or more _parameterized types_ to
1063 appear in its signature. Each type parameter must be explicitly
1064 declared, in an angle-bracket-enclosed, comma-separated list following
1068 fn iter<T>(seq: &[T], f: |T|) {
1069 for elt in seq.iter() { f(elt); }
1071 fn map<T, U>(seq: &[T], f: |T| -> U) -> Vec<U> {
1072 let mut acc = vec![];
1073 for elt in seq.iter() { acc.push(f(elt)); }
1078 Inside the function signature and body, the name of the type parameter
1079 can be used as a type name.
1081 When a generic function is referenced, its type is instantiated based
1082 on the context of the reference. For example, calling the `iter`
1083 function defined above on `[1, 2]` will instantiate type parameter `T`
1084 with `int`, and require the closure parameter to have type
1087 The type parameters can also be explicitly supplied in a trailing
1088 [path](#paths) component after the function name. This might be necessary
1089 if there is not sufficient context to determine the type parameters. For
1090 example, `mem::size_of::<u32>() == 4`.
1092 Since a parameter type is opaque to the generic function, the set of
1093 operations that can be performed on it is limited. Values of parameter
1094 type can only be moved, not copied.
1097 fn id<T>(x: T) -> T { x }
1100 Similarly, [trait](#traits) bounds can be specified for type
1101 parameters to allow methods with that trait to be called on values
1107 Unsafe operations are those that potentially violate the memory-safety guarantees of Rust's static semantics.
1109 The following language level features cannot be used in the safe subset of Rust:
1111 - Dereferencing a [raw pointer](#pointer-types).
1112 - Reading or writing a [mutable static variable](#mutable-statics).
1113 - Calling an unsafe function (including an intrinsic or foreign function).
1115 ##### Unsafe functions
1117 Unsafe functions are functions that are not safe in all contexts and/or for all possible inputs.
1118 Such a function must be prefixed with the keyword `unsafe`.
1122 A block of code can also be prefixed with the `unsafe` keyword, to permit calling `unsafe` functions
1123 or dereferencing raw pointers within a safe function.
1125 When a programmer has sufficient conviction that a sequence of potentially unsafe operations is
1126 actually safe, they can encapsulate that sequence (taken as a whole) within an `unsafe` block. The
1127 compiler will consider uses of such code safe, in the surrounding context.
1129 Unsafe blocks are used to wrap foreign libraries, make direct use of hardware or implement features
1130 not directly present in the language. For example, Rust provides the language features necessary to
1131 implement memory-safe concurrency in the language but the implementation of tasks and message
1132 passing is in the standard library.
1134 Rust's type system is a conservative approximation of the dynamic safety requirements, so in some
1135 cases there is a performance cost to using safe code. For example, a doubly-linked list is not a
1136 tree structure and can only be represented with managed or reference-counted pointers in safe code.
1137 By using `unsafe` blocks to represent the reverse links as raw pointers, it can be implemented with
1138 only owned pointers.
1140 ##### Behavior considered unsafe
1142 This is a list of behavior which is forbidden in all Rust code. Type checking provides the guarantee
1143 that these issues are never caused by safe code. An `unsafe` block or function is responsible for
1144 never invoking this behaviour or exposing an API making it possible for it to occur in safe code.
1147 * Dereferencing a null/dangling raw pointer
1148 * Mutating an immutable value/reference
1149 * Reads of [undef](http://llvm.org/docs/LangRef.html#undefined-values) (uninitialized) memory
1150 * Breaking the [pointer aliasing rules](http://llvm.org/docs/LangRef.html#pointer-aliasing-rules)
1151 with raw pointers (a subset of the rules used by C)
1152 * Invoking undefined behavior via compiler intrinsics:
1153 * Indexing outside of the bounds of an object with `std::ptr::offset` (`offset` intrinsic), with
1154 the exception of one byte past the end which is permitted.
1155 * Using `std::ptr::copy_nonoverlapping_memory` (`memcpy32`/`memcpy64` instrinsics) on
1157 * Invalid values in primitive types, even in private fields/locals:
1158 * Dangling/null pointers in non-raw pointers, or slices
1159 * A value other than `false` (0) or `true` (1) in a `bool`
1160 * A discriminant in an `enum` not included in the type definition
1161 * A value in a `char` which is a surrogate or above `char::MAX`
1162 * non-UTF-8 byte sequences in a `str`
1164 ##### Behaviour not considered unsafe
1166 This is a list of behaviour not considered *unsafe* in Rust terms, but that may be undesired.
1169 * Reading data from private fields (`std::repr`, `format!("{:?}", x)`)
1170 * Leaks due to reference count cycles, even in the global heap
1171 * Exiting without calling destructors
1173 * Accessing/modifying the file system
1174 * Unsigned integer overflow (well-defined as wrapping)
1175 * Signed integer overflow (well-defined as two's complement representation wrapping)
1177 #### Diverging functions
1179 A special kind of function can be declared with a `!` character where the
1180 output slot type would normally be. For example:
1183 fn my_err(s: &str) -> ! {
1189 We call such functions "diverging" because they never return a value to the
1190 caller. Every control path in a diverging function must end with a
1191 `fail!()` or a call to another diverging function on every
1192 control path. The `!` annotation does *not* denote a type. Rather, the result
1193 type of a diverging function is a special type called $\bot$ ("bottom") that
1194 unifies with any type. Rust has no syntax for $\bot$.
1196 It might be necessary to declare a diverging function because as mentioned
1197 previously, the typechecker checks that every control path in a function ends
1198 with a [`return`](#return-expressions) or diverging expression. So, if `my_err`
1199 were declared without the `!` annotation, the following code would not
1203 # fn my_err(s: &str) -> ! { fail!() }
1205 fn f(i: int) -> int {
1210 my_err("Bad number!");
1215 This will not compile without the `!` annotation on `my_err`,
1216 since the `else` branch of the conditional in `f` does not return an `int`,
1217 as required by the signature of `f`.
1218 Adding the `!` annotation to `my_err` informs the typechecker that,
1219 should control ever enter `my_err`, no further type judgments about `f` need to hold,
1220 since control will never resume in any context that relies on those judgments.
1221 Thus the return type on `f` only needs to reflect the `if` branch of the conditional.
1224 #### Extern functions
1226 Extern functions are part of Rust's foreign function interface,
1227 providing the opposite functionality to [external blocks](#external-blocks).
1228 Whereas external blocks allow Rust code to call foreign code,
1229 extern functions with bodies defined in Rust code _can be called by foreign
1230 code_. They are defined in the same way as any other Rust function,
1231 except that they have the `extern` modifier.
1234 // Declares an extern fn, the ABI defaults to "C"
1235 extern fn new_int() -> int { 0 }
1237 // Declares an extern fn with "stdcall" ABI
1238 extern "stdcall" fn new_int_stdcall() -> int { 0 }
1241 Unlike normal functions, extern fns have an `extern "ABI" fn()`.
1242 This is the same type as the functions declared in an extern
1246 # extern fn new_int() -> int { 0 }
1247 let fptr: extern "C" fn() -> int = new_int;
1250 Extern functions may be called directly from Rust code as Rust uses large,
1251 contiguous stack segments like C.
1253 ### Type definitions
1255 A _type definition_ defines a new name for an existing [type](#types). Type
1256 definitions are declared with the keyword `type`. Every value has a single,
1257 specific type; the type-specified aspects of a value include:
1259 * Whether the value is composed of sub-values or is indivisible.
1260 * Whether the value represents textual or numerical information.
1261 * Whether the value represents integral or floating-point information.
1262 * The sequence of memory operations required to access the value.
1263 * The [kind](#type-kinds) of the type.
1265 For example, the type `(u8, u8)` defines the set of immutable values that are composite pairs,
1266 each containing two unsigned 8-bit integers accessed by pattern-matching and laid out in memory with the `x` component preceding the `y` component.
1270 A _structure_ is a nominal [structure type](#structure-types) defined with the keyword `struct`.
1272 An example of a `struct` item and its use:
1275 struct Point {x: int, y: int}
1276 let p = Point {x: 10, y: 11};
1280 A _tuple structure_ is a nominal [tuple type](#tuple-types), also defined with the keyword `struct`.
1284 struct Point(int, int);
1285 let p = Point(10, 11);
1286 let px: int = match p { Point(x, _) => x };
1289 A _unit-like struct_ is a structure without any fields, defined by leaving off the list of fields entirely.
1290 Such types will have a single value, just like the [unit value `()`](#unit-and-boolean-literals) of the unit type.
1295 let c = [Cookie, Cookie, Cookie, Cookie];
1298 By using the `struct_inherit` feature gate, structures may use single inheritance. A Structure may only
1299 inherit from a single other structure, called the _super-struct_. The inheriting structure (sub-struct)
1300 acts as if all fields in the super-struct were present in the sub-struct. Fields declared in a sub-struct
1301 must not have the same name as any field in any (transitive) super-struct. All fields (both declared
1302 and inherited) must be specified in any initializers. Inheritance between structures does not give
1303 subtyping or coercion. The super-struct and sub-struct must be defined in the same crate. The super-struct
1304 must be declared using the `virtual` keyword.
1308 virtual struct Sup { x: int }
1309 struct Sub : Sup { y: int }
1310 let s = Sub {x: 10, y: 11};
1316 An _enumeration_ is a simultaneous definition of a nominal [enumerated type](#enumerated-types) as well as a set of *constructors*,
1317 that can be used to create or pattern-match values of the corresponding enumerated type.
1319 Enumerations are declared with the keyword `enum`.
1321 An example of an `enum` item and its use:
1329 let mut a: Animal = Dog;
1333 Enumeration constructors can have either named or unnamed fields:
1336 # #![feature(struct_variant)]
1340 Cat { name: String, weight: f64 }
1343 let mut a: Animal = Dog("Cocoa".to_string(), 37.2);
1344 a = Cat { name: "Spotty".to_string(), weight: 2.7 };
1348 In this example, `Cat` is a _struct-like enum variant_,
1349 whereas `Dog` is simply called an enum variant.
1354 static_item : "static" ident ':' type '=' expr ';' ;
1357 A *static item* is a named _constant value_ stored in the global data section of a crate.
1358 Immutable static items are stored in the read-only data section.
1359 The constant value bound to a static item is, like all constant values, evaluated at compile time.
1360 Static items have the `static` lifetime, which outlives all other lifetimes in a Rust program.
1361 Static items are declared with the `static` keyword.
1362 A static item must have a _constant expression_ giving its definition.
1364 Static items must be explicitly typed.
1365 The type may be ```bool```, ```char```, a number, or a type derived from those primitive types.
1366 The derived types are references with the `static` lifetime,
1367 fixed-size arrays, tuples, and structs.
1370 static BIT1: uint = 1 << 0;
1371 static BIT2: uint = 1 << 1;
1373 static BITS: [uint, ..2] = [BIT1, BIT2];
1374 static STRING: &'static str = "bitstring";
1376 struct BitsNStrings<'a> {
1377 mybits: [uint, ..2],
1381 static bits_n_strings: BitsNStrings<'static> = BitsNStrings {
1387 #### Mutable statics
1389 If a static item is declared with the ```mut``` keyword, then it is allowed to
1390 be modified by the program. One of Rust's goals is to make concurrency bugs hard
1391 to run into, and this is obviously a very large source of race conditions or
1392 other bugs. For this reason, an ```unsafe``` block is required when either
1393 reading or writing a mutable static variable. Care should be taken to ensure
1394 that modifications to a mutable static are safe with respect to other tasks
1395 running in the same process.
1397 Mutable statics are still very useful, however. They can be used with C
1398 libraries and can also be bound from C libraries (in an ```extern``` block).
1401 # fn atomic_add(_: &mut uint, _: uint) -> uint { 2 }
1403 static mut LEVELS: uint = 0;
1405 // This violates the idea of no shared state, and this doesn't internally
1406 // protect against races, so this function is `unsafe`
1407 unsafe fn bump_levels_unsafe1() -> uint {
1413 // Assuming that we have an atomic_add function which returns the old value,
1414 // this function is "safe" but the meaning of the return value may not be what
1415 // callers expect, so it's still marked as `unsafe`
1416 unsafe fn bump_levels_unsafe2() -> uint {
1417 return atomic_add(&mut LEVELS, 1);
1423 A _trait_ describes a set of method types.
1425 Traits can include default implementations of methods,
1426 written in terms of some unknown [`self` type](#self-types);
1427 the `self` type may either be completely unspecified,
1428 or constrained by some other trait.
1430 Traits are implemented for specific types through separate [implementations](#implementations).
1433 # type Surface = int;
1434 # type BoundingBox = int;
1436 fn draw(&self, Surface);
1437 fn bounding_box(&self) -> BoundingBox;
1441 This defines a trait with two methods.
1442 All values that have [implementations](#implementations) of this trait in scope can have their `draw` and `bounding_box` methods called,
1443 using `value.bounding_box()` [syntax](#method-call-expressions).
1445 Type parameters can be specified for a trait to make it generic.
1446 These appear after the trait name, using the same syntax used in [generic functions](#generic-functions).
1450 fn len(&self) -> uint;
1451 fn elt_at(&self, n: uint) -> T;
1452 fn iter(&self, |T|);
1456 Generic functions may use traits as _bounds_ on their type parameters.
1457 This will have two effects: only types that have the trait may instantiate the parameter,
1458 and within the generic function,
1459 the methods of the trait can be called on values that have the parameter's type.
1463 # type Surface = int;
1464 # trait Shape { fn draw(&self, Surface); }
1465 fn draw_twice<T: Shape>(surface: Surface, sh: T) {
1471 Traits also define an [object type](#object-types) with the same name as the trait.
1472 Values of this type are created by [casting](#type-cast-expressions) pointer values
1473 (pointing to a type for which an implementation of the given trait is in scope)
1474 to pointers to the trait name, used as a type.
1478 # impl Shape for int { }
1479 # let mycircle = 0i;
1480 let myshape: Box<Shape> = box mycircle as Box<Shape>;
1483 The resulting value is a box containing the value that was cast,
1484 along with information that identifies the methods of the implementation that was used.
1485 Values with a trait type can have [methods called](#method-call-expressions) on them,
1486 for any method in the trait,
1487 and can be used to instantiate type parameters that are bounded by the trait.
1489 Trait methods may be static,
1490 which means that they lack a `self` argument.
1491 This means that they can only be called with function call syntax (`f(x)`)
1492 and not method call syntax (`obj.f()`).
1493 The way to refer to the name of a static method is to qualify it with the trait name,
1494 treating the trait name like a module.
1499 fn from_int(n: int) -> Self;
1502 fn from_int(n: int) -> f64 { n as f64 }
1504 let x: f64 = Num::from_int(42);
1507 Traits may inherit from other traits. For example, in
1510 trait Shape { fn area() -> f64; }
1511 trait Circle : Shape { fn radius() -> f64; }
1514 the syntax `Circle : Shape` means that types that implement `Circle` must also have an implementation for `Shape`.
1515 Multiple supertraits are separated by `+`, `trait Circle : Shape + PartialEq { }`.
1516 In an implementation of `Circle` for a given type `T`, methods can refer to `Shape` methods,
1517 since the typechecker checks that any type with an implementation of `Circle` also has an implementation of `Shape`.
1519 In type-parameterized functions,
1520 methods of the supertrait may be called on values of subtrait-bound type parameters.
1521 Referring to the previous example of `trait Circle : Shape`:
1524 # trait Shape { fn area(&self) -> f64; }
1525 # trait Circle : Shape { fn radius(&self) -> f64; }
1526 fn radius_times_area<T: Circle>(c: T) -> f64 {
1527 // `c` is both a Circle and a Shape
1528 c.radius() * c.area()
1532 Likewise, supertrait methods may also be called on trait objects.
1535 # trait Shape { fn area(&self) -> f64; }
1536 # trait Circle : Shape { fn radius(&self) -> f64; }
1537 # impl Shape for int { fn area(&self) -> f64 { 0.0 } }
1538 # impl Circle for int { fn radius(&self) -> f64 { 0.0 } }
1540 let mycircle: Circle = ~mycircle as ~Circle;
1541 let nonsense = mycircle.radius() * mycircle.area();
1546 An _implementation_ is an item that implements a [trait](#traits) for a specific type.
1548 Implementations are defined with the keyword `impl`.
1551 # struct Point {x: f64, y: f64};
1552 # type Surface = int;
1553 # struct BoundingBox {x: f64, y: f64, width: f64, height: f64};
1554 # trait Shape { fn draw(&self, Surface); fn bounding_box(&self) -> BoundingBox; }
1555 # fn do_draw_circle(s: Surface, c: Circle) { }
1561 impl Shape for Circle {
1562 fn draw(&self, s: Surface) { do_draw_circle(s, *self); }
1563 fn bounding_box(&self) -> BoundingBox {
1564 let r = self.radius;
1565 BoundingBox{x: self.center.x - r, y: self.center.y - r,
1566 width: 2.0 * r, height: 2.0 * r}
1571 It is possible to define an implementation without referring to a trait.
1572 The methods in such an implementation can only be used
1573 as direct calls on the values of the type that the implementation targets.
1574 In such an implementation, the trait type and `for` after `impl` are omitted.
1575 Such implementations are limited to nominal types (enums, structs),
1576 and the implementation must appear in the same module or a sub-module as the `self` type.
1578 When a trait _is_ specified in an `impl`,
1579 all methods declared as part of the trait must be implemented,
1580 with matching types and type parameter counts.
1582 An implementation can take type parameters,
1583 which can be different from the type parameters taken by the trait it implements.
1584 Implementation parameters are written after the `impl` keyword.
1588 impl<T> Seq<T> for Vec<T> {
1591 impl Seq<bool> for u32 {
1592 /* Treat the integer as a sequence of bits */
1599 extern_block_item : "extern" '{' extern_block '}' ;
1600 extern_block : [ foreign_fn ] * ;
1603 External blocks form the basis for Rust's foreign function interface.
1604 Declarations in an external block describe symbols
1605 in external, non-Rust libraries.
1607 Functions within external blocks
1608 are declared in the same way as other Rust functions,
1609 with the exception that they may not have a body
1610 and are instead terminated by a semicolon.
1614 use libc::{c_char, FILE};
1617 fn fopen(filename: *c_char, mode: *c_char) -> *FILE;
1622 Functions within external blocks may be called by Rust code,
1623 just like functions defined in Rust.
1624 The Rust compiler automatically translates
1625 between the Rust ABI and the foreign ABI.
1627 A number of [attributes](#attributes) control the behavior of external
1630 By default external blocks assume that the library they are calling
1631 uses the standard C "cdecl" ABI. Other ABIs may be specified using
1632 an `abi` string, as shown here:
1635 // Interface to the Windows API
1636 extern "stdcall" { }
1639 The `link` attribute allows the name of the library to be specified. When
1640 specified the compiler will attempt to link against the native library of the
1644 #[link(name = "crypto")]
1648 The type of a function declared in an extern block is `extern "abi" fn(A1,
1649 ..., An) -> R`, where `A1...An` are the declared types of its arguments and
1650 `R` is the declared return type.
1652 ## Visibility and Privacy
1654 These two terms are often used interchangeably, and what they are attempting to
1655 convey is the answer to the question "Can this item be used at this location?"
1657 Rust's name resolution operates on a global hierarchy of namespaces. Each level
1658 in the hierarchy can be thought of as some item. The items are one of those
1659 mentioned above, but also include external crates. Declaring or defining a new
1660 module can be thought of as inserting a new tree into the hierarchy at the
1661 location of the definition.
1663 To control whether interfaces can be used across modules, Rust checks each use
1664 of an item to see whether it should be allowed or not. This is where privacy
1665 warnings are generated, or otherwise "you used a private item of another module
1666 and weren't allowed to."
1668 By default, everything in rust is *private*, with one exception. Enum variants
1669 in a `pub` enum are also public by default. You are allowed to alter this
1670 default visibility with the `priv` keyword. When an item is declared as `pub`,
1671 it can be thought of as being accessible to the outside world. For example:
1675 // Declare a private struct
1678 // Declare a public struct with a private field
1683 // Declare a public enum with two public variants
1685 PubliclyAccessibleState,
1686 PubliclyAccessibleState2,
1690 With the notion of an item being either public or private, Rust allows item
1691 accesses in two cases:
1693 1. If an item is public, then it can be used externally through any of its
1695 2. If an item is private, it may be accessed by the current module and its
1698 These two cases are surprisingly powerful for creating module hierarchies
1699 exposing public APIs while hiding internal implementation details. To help
1700 explain, here's a few use cases and what they would entail.
1702 * A library developer needs to expose functionality to crates which link against
1703 their library. As a consequence of the first case, this means that anything
1704 which is usable externally must be `pub` from the root down to the destination
1705 item. Any private item in the chain will disallow external accesses.
1707 * A crate needs a global available "helper module" to itself, but it doesn't
1708 want to expose the helper module as a public API. To accomplish this, the root
1709 of the crate's hierarchy would have a private module which then internally has
1710 a "public api". Because the entire crate is a descendant of the root, then the
1711 entire local crate can access this private module through the second case.
1713 * When writing unit tests for a module, it's often a common idiom to have an
1714 immediate child of the module to-be-tested named `mod test`. This module could
1715 access any items of the parent module through the second case, meaning that
1716 internal implementation details could also be seamlessly tested from the child
1719 In the second case, it mentions that a private item "can be accessed" by the
1720 current module and its descendants, but the exact meaning of accessing an item
1721 depends on what the item is. Accessing a module, for example, would mean looking
1722 inside of it (to import more items). On the other hand, accessing a function
1723 would mean that it is invoked. Additionally, path expressions and import
1724 statements are considered to access an item in the sense that the
1725 import/expression is only valid if the destination is in the current visibility
1728 Here's an example of a program which exemplifies the three cases outlined above.
1731 // This module is private, meaning that no external crate can access this
1732 // module. Because it is private at the root of this current crate, however, any
1733 // module in the crate may access any publicly visible item in this module.
1734 mod crate_helper_module {
1736 // This function can be used by anything in the current crate
1737 pub fn crate_helper() {}
1739 // This function *cannot* be used by anything else in the crate. It is not
1740 // publicly visible outside of the `crate_helper_module`, so only this
1741 // current module and its descendants may access it.
1742 fn implementation_detail() {}
1745 // This function is "public to the root" meaning that it's available to external
1746 // crates linking against this one.
1747 pub fn public_api() {}
1749 // Similarly to 'public_api', this module is public so external crates may look
1752 use crate_helper_module;
1754 pub fn my_method() {
1755 // Any item in the local crate may invoke the helper module's public
1756 // interface through a combination of the two rules above.
1757 crate_helper_module::crate_helper();
1760 // This function is hidden to any module which is not a descendant of
1762 fn my_implementation() {}
1768 fn test_my_implementation() {
1769 // Because this module is a descendant of `submodule`, it's allowed
1770 // to access private items inside of `submodule` without a privacy
1772 super::my_implementation();
1780 For a rust program to pass the privacy checking pass, all paths must be valid
1781 accesses given the two rules above. This includes all use statements,
1782 expressions, types, etc.
1784 ### Re-exporting and Visibility
1786 Rust allows publicly re-exporting items through a `pub use` directive. Because
1787 this is a public directive, this allows the item to be used in the current
1788 module through the rules above. It essentially allows public access into the
1789 re-exported item. For example, this program is valid:
1792 pub use api = self::implementation;
1794 mod implementation {
1801 This means that any external crate referencing `implementation::f` would receive
1802 a privacy violation, while the path `api::f` would be allowed.
1804 When re-exporting a private item, it can be thought of as allowing the "privacy
1805 chain" being short-circuited through the reexport instead of passing through the
1806 namespace hierarchy as it normally would.
1808 ### Glob imports and Visibility
1810 Currently glob imports are considered an "experimental" language feature. For
1811 sanity purpose along with helping the implementation, glob imports will only
1812 import public items from their destination, not private items.
1814 > **Note:** This is subject to change, glob exports may be removed entirely or
1815 > they could possibly import private items for a privacy error to later be
1816 > issued if the item is used.
1821 attribute : '#' '!' ? '[' meta_item ']' ;
1822 meta_item : ident [ '=' literal
1823 | '(' meta_seq ')' ] ? ;
1824 meta_seq : meta_item [ ',' meta_seq ] ? ;
1827 Static entities in Rust — crates, modules and items — may have _attributes_
1828 applied to them. Attributes in Rust are modeled on Attributes in ECMA-335,
1829 with the syntax coming from ECMA-334 (C#). An attribute is a general,
1830 free-form metadatum that is interpreted according to name, convention, and
1831 language and compiler version. Attributes may appear as any of:
1833 * A single identifier, the attribute name
1834 * An identifier followed by the equals sign '=' and a literal, providing a
1836 * An identifier followed by a parenthesized list of sub-attribute arguments
1838 Attributes with a bang ("!") after the hash ("#") apply to the item that the
1839 attribute is declared within. Attributes that do not have a bang after the
1840 hash apply to the item that follows the attribute.
1842 An example of attributes:
1845 // General metadata applied to the enclosing module or crate.
1848 // A function marked as a unit test
1854 // A conditionally-compiled module
1855 #[cfg(target_os="linux")]
1860 // A lint attribute used to suppress a warning/error
1861 #[allow(non_camel_case_types)]
1865 > **Note:** At some point in the future, the compiler will distinguish between
1866 > language-reserved and user-available attributes. Until then, there is
1867 > effectively no difference between an attribute handled by a loadable syntax
1868 > extension and the compiler.
1870 ### Crate-only attributes
1872 - `crate_id` - specify the this crate's crate ID.
1873 - `crate_type` - see [linkage](#linkage).
1874 - `feature` - see [compiler features](#compiler-features).
1875 - `no_main` - disable emitting the `main` symbol. Useful when some other
1876 object being linked to defines `main`.
1877 - `no_start` - disable linking to the `native` crate, which specifies the
1878 "start" language item.
1879 - `no_std` - disable linking to the `std` crate.
1880 - `no_builtins` - disable optimizing certain code patterns to invocations of
1881 library functions that are assumed to exist
1883 ### Module-only attributes
1885 - `macro_escape` - macros defined in this module will be visible in the
1886 module's parent, after this module has been included.
1887 - `no_implicit_prelude` - disable injecting `use std::prelude::*` in this
1889 - `path` - specifies the file to load the module from. `#[path="foo.rs"] mod
1890 bar;` is equivalent to `mod bar { /* contents of foo.rs */ }`. The path is
1891 taken relative to the directory that the current module is in.
1893 ### Function-only attributes
1895 - `plugin_registrar` - mark this function as the registration point for
1896 compiler plugins, such as loadable syntax extensions.
1897 - `main` - indicates that this function should be passed to the entry point,
1898 rather than the function in the crate root named `main`.
1899 - `start` - indicates that this function should be used as the entry point,
1900 overriding the "start" language item. See the "start" [language
1901 item](#language-items) for more details.
1903 ### Static-only attributes
1905 - `thread_local` - on a `static mut`, this signals that the value of this
1906 static may change depending on the current thread. The exact consequences of
1907 this are implementation-defined.
1911 On an `extern` block, the following attributes are interpreted:
1913 - `link_args` - specify arguments to the linker, rather than just the library
1914 name and type. This is feature gated and the exact behavior is
1915 implementation-defined (due to variety of linker invocation syntax).
1916 - `link` - indicate that a native library should be linked to for the
1917 declarations in this block to be linked correctly. See [external
1918 blocks](#external-blocks)
1920 On declarations inside an `extern` block, the following attributes are
1923 - `link_name` - the name of the symbol that this function or static should be
1925 - `linkage` - on a static, this specifies the [linkage
1926 type](http://llvm.org/docs/LangRef.html#linkage-types).
1928 ### Miscellaneous attributes
1930 - `link_section` - on statics and functions, this specifies the section of the
1931 object file that this item's contents will be placed into.
1932 - `macro_export` - export a macro for cross-crate usage.
1933 - `no_mangle` - on any item, do not apply the standard name mangling. Set the
1934 symbol for this item to its identifier.
1935 - `packed` - on structs or enums, eliminate any padding that would be used to
1937 - `repr` - on C-like enums, this sets the underlying type used for
1938 representation. Useful for FFI. Takes one argument, which is the primitive
1939 type this enum should be represented for, or `C`, which specifies that it
1940 should be the default `enum` size of the C ABI for that platform. Note that
1941 enum representation in C is undefined, and this may be incorrect when the C
1942 code is compiled with certain flags.
1943 - `simd` - on certain tuple structs, derive the arithmetic operators, which
1944 lower to the target's SIMD instructions, if any; the `simd` feature gate
1945 is necessary to use this attribute.
1946 - `static_assert` - on statics whose type is `bool`, terminates compilation
1947 with an error if it is not initialized to `true`.
1948 - `unsafe_destructor` - allow implementations of the "drop" language item
1949 where the type it is implemented for does not implement the "send" language
1950 item; the `unsafe_destructor` feature gate is needed to use this attribute
1951 - `unsafe_no_drop_flag` - on structs, remove the flag that prevents
1952 destructors from being run twice. Destructors might be run multiple times on
1953 the same object with this attribute.
1955 ### Conditional compilation
1957 Sometimes one wants to have different compiler outputs from the same code,
1958 depending on build target, such as targeted operating system, or to enable
1961 There are two kinds of configuration options, one that is either defined or not
1962 (`#[cfg(foo)]`), and the other that contains a string that can be checked
1963 against (`#[cfg(bar = "baz")]` (currently only compiler-defined configuration
1964 options can have the latter form).
1967 // The function is only included in the build when compiling for OSX
1968 #[cfg(target_os = "macos")]
1973 // This function is only included when either foo or bar is defined
1976 fn needs_foo_or_bar() {
1980 // This function is only included when compiling for a unixish OS with a 32-bit
1982 #[cfg(unix, target_word_size = "32")]
1983 fn on_32bit_unix() {
1988 This illustrates some conditional compilation can be achieved using the
1989 `#[cfg(...)]` attribute. Note that `#[cfg(foo, bar)]` is a condition that needs
1990 both `foo` and `bar` to be defined while `#[cfg(foo)] #[cfg(bar)]` only needs
1991 one of `foo` and `bar` to be defined (this resembles in the disjunctive normal
1992 form). Additionally, one can reverse a condition by enclosing it in a
1993 `not(...)`, like e. g. `#[cfg(not(target_os = "win32"))]`.
1995 The following configurations must be defined by the implementation:
1997 * `target_arch = "..."`. Target CPU architecture, such as `"x86"`, `"x86_64"`
1998 `"mips"`, or `"arm"`.
1999 * `target_endian = "..."`. Endianness of the target CPU, either `"little"` or
2001 * `target_family = "..."`. Operating system family of the target, e. g.
2002 `"unix"` or `"windows"`. The value of this configuration option is defined as
2003 a configuration itself, like `unix` or `windows`.
2004 * `target_os = "..."`. Operating system of the target, examples include
2005 `"win32"`, `"macos"`, `"linux"`, `"android"` or `"freebsd"`.
2006 * `target_word_size = "..."`. Target word size in bits. This is set to `"32"`
2007 for targets with 32-bit pointers, and likewise set to `"64"` for 64-bit
2009 * `unix`. See `target_family`.
2010 * `windows`. See `target_family`.
2012 ### Lint check attributes
2014 A lint check names a potentially undesirable coding pattern, such as
2015 unreachable code or omitted documentation, for the static entity to
2016 which the attribute applies.
2018 For any lint check `C`:
2020 * `warn(C)` warns about violations of `C` but continues compilation,
2021 * `deny(C)` signals an error after encountering a violation of `C`,
2022 * `allow(C)` overrides the check for `C` so that violations will go
2024 * `forbid(C)` is the same as `deny(C)`, but also forbids changing the lint
2027 The lint checks supported by the compiler can be found via `rustc -W help`,
2028 along with their default settings.
2032 // Missing documentation is ignored here
2033 #[allow(missing_doc)]
2034 pub fn undocumented_one() -> int { 1 }
2036 // Missing documentation signals a warning here
2037 #[warn(missing_doc)]
2038 pub fn undocumented_too() -> int { 2 }
2040 // Missing documentation signals an error here
2041 #[deny(missing_doc)]
2042 pub fn undocumented_end() -> int { 3 }
2046 This example shows how one can use `allow` and `warn` to toggle
2047 a particular check on and off.
2050 #[warn(missing_doc)]
2052 #[allow(missing_doc)]
2054 // Missing documentation is ignored here
2055 pub fn undocumented_one() -> int { 1 }
2057 // Missing documentation signals a warning here,
2058 // despite the allow above.
2059 #[warn(missing_doc)]
2060 pub fn undocumented_two() -> int { 2 }
2063 // Missing documentation signals a warning here
2064 pub fn undocumented_too() -> int { 3 }
2068 This example shows how one can use `forbid` to disallow uses
2069 of `allow` for that lint check.
2072 #[forbid(missing_doc)]
2074 // Attempting to toggle warning signals an error here
2075 #[allow(missing_doc)]
2077 pub fn undocumented_too() -> int { 2 }
2083 Some primitive Rust operations are defined in Rust code, rather than being
2084 implemented directly in C or assembly language. The definitions of these
2085 operations have to be easy for the compiler to find. The `lang` attribute
2086 makes it possible to declare these operations. For example, the `str` module
2087 in the Rust standard library defines the string equality function:
2091 pub fn eq_slice(a: &str, b: &str) -> bool {
2096 The name `str_eq` has a special meaning to the Rust compiler,
2097 and the presence of this definition means that it will use this definition
2098 when generating calls to the string equality function.
2100 A complete list of the built-in language items follows:
2102 #### Built-in Traits
2105 : Able to be sent across task boundaries.
2107 : Has a size known at compile time.
2109 : Types that do not move ownership when used by-value.
2111 : Able to be safely shared between tasks when aliased.
2117 These language items are traits:
2120 : Elements can be added (for example, integers and floats).
2122 : Elements can be subtracted.
2124 : Elements can be multiplied.
2126 : Elements have a division operation.
2128 : Elements have a remainder operation.
2130 : Elements can be negated arithmetically.
2132 : Elements can be negated logically.
2134 : Elements have an exclusive-or operation.
2136 : Elements have a bitwise `and` operation.
2138 : Elements have a bitwise `or` operation.
2140 : Elements have a left shift operation.
2142 : Elements have a right shift operation.
2144 : Elements can be indexed.
2146 : Elements can be compared for equality.
2148 : Elements have a partial ordering.
2150 : `*` can be applied, yielding a reference to another type
2152 : `*` can be applied, yielding a mutable reference to another type
2155 These are functions:
2158 : Compare two strings (`&str`) for equality.
2160 : Compare two owned strings (`String`) for equality.
2162 : Return a new unique string
2163 containing a copy of the contents of a unique string.
2168 : A type whose contents can be mutated through an immutable reference
2170 : The type returned by the `type_id` intrinsic.
2174 These types help drive the compiler's analysis
2177 : The type parameter should be considered covariant
2178 * `contravariant_type`
2179 : The type parameter should be considered contravariant
2181 : The type parameter should be considered invariant
2182 * `covariant_lifetime`
2183 : The lifetime parameter should be considered covariant
2184 * `contravariant_lifetime`
2185 : The lifetime parameter should be considered contravariant
2186 * `invariant_lifetime`
2187 : The lifetime parameter should be considered invariant
2189 : This type does not implement "send", even if eligible
2191 : This type does not implement "copy", even if eligible
2193 : This type does not implement "share", even if eligible
2195 : This type implements "managed"
2198 : Abort the program with an error.
2199 * `fail_bounds_check`
2200 : Abort the program with a bounds check error.
2202 : Allocate memory on the exchange heap.
2204 : Free memory that was allocated on the exchange heap.
2206 : Allocate memory on the managed heap.
2208 : Free memory that was allocated on the managed heap.
2210 > **Note:** This list is likely to become out of date. We should auto-generate it
2211 > from `librustc/middle/lang_items.rs`.
2213 ### Inline attributes
2215 The inline attribute is used to suggest to the compiler to perform an inline
2216 expansion and place a copy of the function or static in the caller rather than
2217 generating code to call the function or access the static where it is defined.
2219 The compiler automatically inlines functions based on internal heuristics.
2220 Incorrectly inlining functions can actually making the program slower, so it
2221 should be used with care.
2223 Immutable statics are always considered inlineable
2224 unless marked with `#[inline(never)]`.
2226 whether two different inlineable statics
2227 have the same memory address.
2229 the compiler is free
2230 to collapse duplicate inlineable statics together.
2232 `#[inline]` and `#[inline(always)]` always causes the function to be serialized
2233 into crate metadata to allow cross-crate inlining.
2235 There are three different types of inline attributes:
2237 * `#[inline]` hints the compiler to perform an inline expansion.
2238 * `#[inline(always)]` asks the compiler to always perform an inline expansion.
2239 * `#[inline(never)]` asks the compiler to never perform an inline expansion.
2243 The `deriving` attribute allows certain traits to be automatically
2244 implemented for data structures. For example, the following will
2245 create an `impl` for the `PartialEq` and `Clone` traits for `Foo`, the type
2246 parameter `T` will be given the `PartialEq` or `Clone` constraints for the
2250 #[deriving(PartialEq, Clone)]
2257 The generated `impl` for `PartialEq` is equivalent to
2260 # struct Foo<T> { a: int, b: T }
2261 impl<T: PartialEq> PartialEq for Foo<T> {
2262 fn eq(&self, other: &Foo<T>) -> bool {
2263 self.a == other.a && self.b == other.b
2266 fn ne(&self, other: &Foo<T>) -> bool {
2267 self.a != other.a || self.b != other.b
2272 Supported traits for `deriving` are:
2274 * Comparison traits: `PartialEq`, `Eq`, `PartialOrd`, `Ord`.
2275 * Serialization: `Encodable`, `Decodable`. These require `serialize`.
2276 * `Clone`, to create `T` from `&T` via a copy.
2277 * `Hash`, to iterate over the bytes in a data type.
2278 * `Rand`, to create a random instance of a data type.
2279 * `Default`, to create an empty instance of a data type.
2280 * `Zero`, to create a zero instance of a numeric data type.
2281 * `FromPrimitive`, to create an instance from a numeric primitive.
2282 * `Show`, to format a value using the `{}` formatter.
2286 One can indicate the stability of an API using the following attributes:
2288 * `deprecated`: This item should no longer be used, e.g. it has been
2289 replaced. No guarantee of backwards-compatibility.
2290 * `experimental`: This item was only recently introduced or is
2291 otherwise in a state of flux. It may change significantly, or even
2292 be removed. No guarantee of backwards-compatibility.
2293 * `unstable`: This item is still under development, but requires more
2294 testing to be considered stable. No guarantee of backwards-compatibility.
2295 * `stable`: This item is considered stable, and will not change
2296 significantly. Guarantee of backwards-compatibility.
2297 * `frozen`: This item is very stable, and is unlikely to
2298 change. Guarantee of backwards-compatibility.
2299 * `locked`: This item will never change unless a serious bug is
2300 found. Guarantee of backwards-compatibility.
2302 These levels are directly inspired by
2303 [Node.js' "stability index"](http://nodejs.org/api/documentation.html).
2305 Stability levels are inherited, so an items's stability attribute is the
2306 default stability for everything nested underneath it.
2308 There are lints for disallowing items marked with certain levels: `deprecated`,
2309 `experimental` and `unstable`. For now, only `deprecated` warns by default, but
2310 this will change once the standard library has been stabilized.
2311 Stability levels are meant to be promises at the crate
2312 level, so these lints only apply when referencing
2313 items from an _external_ crate, not to items defined within the
2314 current crate. Items with no stability level are considered
2315 to be unstable for the purposes of the lint. One can give an optional
2316 string that will be displayed when the lint flags the use of an item.
2318 For example, if we define one crate called `stability_levels`:
2321 #[deprecated="replaced by `best`"]
2323 // delete everything
2327 // delete fewer things
2336 then the lints will work as follows for a client crate:
2340 extern crate stability_levels;
2341 use stability_levels::{bad, better, best};
2344 bad(); // "warning: use of deprecated item: replaced by `best`"
2346 better(); // "warning: use of unmarked item"
2348 best(); // no warning
2352 > **Note:** Currently these are only checked when applied to
2353 > individual functions, structs, methods and enum variants, *not* to
2354 > entire modules, traits, impls or enums themselves.
2356 ### Compiler Features
2358 Certain aspects of Rust may be implemented in the compiler, but they're not
2359 necessarily ready for every-day use. These features are often of "prototype
2360 quality" or "almost production ready", but may not be stable enough to be
2361 considered a full-fledged language feature.
2363 For this reason, Rust recognizes a special crate-level attribute of the form:
2366 #![feature(feature1, feature2, feature3)]
2369 This directive informs the compiler that the feature list: `feature1`,
2370 `feature2`, and `feature3` should all be enabled. This is only recognized at a
2371 crate-level, not at a module-level. Without this directive, all features are
2372 considered off, and using the features will result in a compiler error.
2374 The currently implemented features of the reference compiler are:
2376 * `macro_rules` - The definition of new macros. This does not encompass
2377 macro-invocation, that is always enabled by default, this only
2378 covers the definition of new macros. There are currently
2379 various problems with invoking macros, how they interact with
2380 their environment, and possibly how they are used outside of
2381 location in which they are defined. Macro definitions are
2382 likely to change slightly in the future, so they are currently
2383 hidden behind this feature.
2385 * `globs` - Importing everything in a module through `*`. This is currently a
2386 large source of bugs in name resolution for Rust, and it's not clear
2387 whether this will continue as a feature or not. For these reasons,
2388 the glob import statement has been hidden behind this feature flag.
2390 * `struct_variant` - Structural enum variants (those with named fields). It is
2391 currently unknown whether this style of enum variant is as
2392 fully supported as the tuple-forms, and it's not certain
2393 that this style of variant should remain in the language.
2394 For now this style of variant is hidden behind a feature
2397 * `once_fns` - Onceness guarantees a closure is only executed once. Defining a
2398 closure as `once` is unlikely to be supported going forward. So
2399 they are hidden behind this feature until they are to be removed.
2401 * `managed_boxes` - Usage of `@` pointers is gated due to many
2402 planned changes to this feature. In the past, this has meant
2403 "a GC pointer", but the current implementation uses
2404 reference counting and will likely change drastically over
2405 time. Additionally, the `@` syntax will no longer be used to
2408 * `asm` - The `asm!` macro provides a means for inline assembly. This is often
2409 useful, but the exact syntax for this feature along with its semantics
2410 are likely to change, so this macro usage must be opted into.
2412 * `non_ascii_idents` - The compiler supports the use of non-ascii identifiers,
2413 but the implementation is a little rough around the
2414 edges, so this can be seen as an experimental feature for
2415 now until the specification of identifiers is fully
2418 * `thread_local` - The usage of the `#[thread_local]` attribute is experimental
2419 and should be seen as unstable. This attribute is used to
2420 declare a `static` as being unique per-thread leveraging
2421 LLVM's implementation which works in concert with the kernel
2422 loader and dynamic linker. This is not necessarily available
2423 on all platforms, and usage of it is discouraged (rust
2424 focuses more on task-local data instead of thread-local
2427 * `link_args` - This attribute is used to specify custom flags to the linker,
2428 but usage is strongly discouraged. The compiler's usage of the
2429 system linker is not guaranteed to continue in the future, and
2430 if the system linker is not used then specifying custom flags
2431 doesn't have much meaning.
2433 If a feature is promoted to a language feature, then all existing programs will
2434 start to receive compilation warnings about #[feature] directives which enabled
2435 the new feature (because the directive is no longer necessary). However, if
2436 a feature is decided to be removed from the language, errors will be issued (if
2437 there isn't a parser error first). The directive in this case is no longer
2438 necessary, and it's likely that existing code will break if the feature isn't
2441 If a unknown feature is found in a directive, it results in a compiler error. An
2442 unknown feature is one which has never been recognized by the compiler.
2444 # Statements and expressions
2446 Rust is _primarily_ an expression language. This means that most forms of
2447 value-producing or effect-causing evaluation are directed by the uniform
2448 syntax category of _expressions_. Each kind of expression can typically _nest_
2449 within each other kind of expression, and rules for evaluation of expressions
2450 involve specifying both the value produced by the expression and the order in
2451 which its sub-expressions are themselves evaluated.
2453 In contrast, statements in Rust serve _mostly_ to contain and explicitly
2454 sequence expression evaluation.
2458 A _statement_ is a component of a block, which is in turn a component of an
2459 outer [expression](#expressions) or [function](#functions).
2461 Rust has two kinds of statement:
2462 [declaration statements](#declaration-statements) and
2463 [expression statements](#expression-statements).
2465 ### Declaration statements
2467 A _declaration statement_ is one that introduces one or more *names* into the enclosing statement block.
2468 The declared names may denote new slots or new items.
2470 #### Item declarations
2472 An _item declaration statement_ has a syntactic form identical to an
2473 [item](#items) declaration within a module. Declaring an item — a function,
2474 enumeration, structure, type, static, trait, implementation or module — locally
2475 within a statement block is simply a way of restricting its scope to a narrow
2476 region containing all of its uses; it is otherwise identical in meaning to
2477 declaring the item outside the statement block.
2479 Note: there is no implicit capture of the function's dynamic environment when
2480 declaring a function-local item.
2482 #### Slot declarations
2485 let_decl : "let" pat [':' type ] ? [ init ] ? ';' ;
2486 init : [ '=' ] expr ;
2489 A _slot declaration_ introduces a new set of slots, given by a pattern.
2490 The pattern may be followed by a type annotation, and/or an initializer expression.
2491 When no type annotation is given, the compiler will infer the type,
2492 or signal an error if insufficient type information is available for definite inference.
2493 Any slots introduced by a slot declaration are visible from the point of declaration until the end of the enclosing block scope.
2495 ### Expression statements
2497 An _expression statement_ is one that evaluates an [expression](#expressions)
2498 and ignores its result.
2499 The type of an expression statement `e;` is always `()`, regardless of the type of `e`.
2500 As a rule, an expression statement's purpose is to trigger the effects of evaluating its expression.
2504 An expression may have two roles: it always produces a *value*, and it may have *effects*
2505 (otherwise known as "side effects").
2506 An expression *evaluates to* a value, and has effects during *evaluation*.
2507 Many expressions contain sub-expressions (operands).
2508 The meaning of each kind of expression dictates several things:
2509 * Whether or not to evaluate the sub-expressions when evaluating the expression
2510 * The order in which to evaluate the sub-expressions
2511 * How to combine the sub-expressions' values to obtain the value of the expression.
2513 In this way, the structure of expressions dictates the structure of execution.
2514 Blocks are just another kind of expression,
2515 so blocks, statements, expressions, and blocks again can recursively nest inside each other
2516 to an arbitrary depth.
2518 #### Lvalues, rvalues and temporaries
2520 Expressions are divided into two main categories: _lvalues_ and _rvalues_.
2521 Likewise within each expression, sub-expressions may occur in _lvalue context_ or _rvalue context_.
2522 The evaluation of an expression depends both on its own category and the context it occurs within.
2524 An lvalue is an expression that represents a memory location. These
2525 expressions are [paths](#path-expressions) (which refer to local
2526 variables, function and method arguments, or static variables),
2527 dereferences (`*expr`), [indexing expressions](#index-expressions)
2528 (`expr[expr]`), and [field references](#field-expressions) (`expr.f`).
2529 All other expressions are rvalues.
2531 The left operand of an [assignment](#assignment-expressions) or
2532 [compound-assignment](#compound-assignment-expressions) expression is an lvalue context,
2533 as is the single operand of a unary [borrow](#unary-operator-expressions).
2534 All other expression contexts are rvalue contexts.
2536 When an lvalue is evaluated in an _lvalue context_, it denotes a memory location;
2537 when evaluated in an _rvalue context_, it denotes the value held _in_ that memory location.
2539 When an rvalue is used in lvalue context, a temporary un-named lvalue is created and used instead.
2540 A temporary's lifetime equals the largest lifetime of any reference that points to it.
2542 #### Moved and copied types
2544 When a [local variable](#memory-slots) is used
2545 as an [rvalue](#lvalues-rvalues-and-temporaries)
2546 the variable will either be moved or copied, depending on its type.
2547 For types that contain [owning pointers](#pointer-types)
2548 or values that implement the special trait `Drop`,
2549 the variable is moved.
2550 All other types are copied.
2552 ### Literal expressions
2554 A _literal expression_ consists of one of the [literal](#literals)
2555 forms described earlier. It directly describes a number, character,
2556 string, boolean value, or the unit value.
2560 "hello"; // string type
2561 '5'; // character type
2565 ### Path expressions
2567 A [path](#paths) used as an expression context denotes either a local variable or an item.
2568 Path expressions are [lvalues](#lvalues-rvalues-and-temporaries).
2570 ### Tuple expressions
2572 Tuples are written by enclosing one or more comma-separated
2573 expressions in parentheses. They are used to create [tuple-typed](#tuple-types)
2582 ### Structure expressions
2585 struct_expr : expr_path '{' ident ':' expr
2586 [ ',' ident ':' expr ] *
2589 [ ',' expr ] * ')' |
2593 There are several forms of structure expressions.
2594 A _structure expression_ consists of the [path](#paths) of a [structure item](#structures),
2595 followed by a brace-enclosed list of one or more comma-separated name-value pairs,
2596 providing the field values of a new instance of the structure.
2597 A field name can be any identifier, and is separated from its value expression by a colon.
2598 The location denoted by a structure field is mutable if and only if the enclosing structure is mutable.
2600 A _tuple structure expression_ consists of the [path](#paths) of a [structure item](#structures),
2601 followed by a parenthesized list of one or more comma-separated expressions
2602 (in other words, the path of a structure item followed by a tuple expression).
2603 The structure item must be a tuple structure item.
2605 A _unit-like structure expression_ consists only of the [path](#paths) of a [structure item](#structures).
2607 The following are examples of structure expressions:
2610 # struct Point { x: f64, y: f64 }
2611 # struct TuplePoint(f64, f64);
2612 # mod game { pub struct User<'a> { pub name: &'a str, pub age: uint, pub score: uint } }
2613 # struct Cookie; fn some_fn<T>(t: T) {}
2614 Point {x: 10.0, y: 20.0};
2615 TuplePoint(10.0, 20.0);
2616 let u = game::User {name: "Joe", age: 35, score: 100_000};
2617 some_fn::<Cookie>(Cookie);
2620 A structure expression forms a new value of the named structure type.
2621 Note that for a given *unit-like* structure type, this will always be the same value.
2623 A structure expression can terminate with the syntax `..` followed by an expression to denote a functional update.
2624 The expression following `..` (the base) must have the same structure type as the new structure type being formed.
2625 The entire expression denotes the result of constructing a new structure
2626 (with the same type as the base expression)
2627 with the given values for the fields that were explicitly specified
2628 and the values in the base expression for all other fields.
2631 # struct Point3d { x: int, y: int, z: int }
2632 let base = Point3d {x: 1, y: 2, z: 3};
2633 Point3d {y: 0, z: 10, .. base};
2636 ### Block expressions
2639 block_expr : '{' [ view_item ] *
2640 [ stmt ';' | item ] *
2644 A _block expression_ is similar to a module in terms of the declarations that
2645 are possible. Each block conceptually introduces a new namespace scope. View
2646 items can bring new names into scopes and declared items are in scope for only
2649 A block will execute each statement sequentially, and then execute the
2650 expression (if given). If the final expression is omitted, the type and return
2651 value of the block are `()`, but if it is provided, the type and return value
2652 of the block are that of the expression itself.
2654 ### Method-call expressions
2657 method_call_expr : expr '.' ident paren_expr_list ;
2660 A _method call_ consists of an expression followed by a single dot, an identifier, and a parenthesized expression-list.
2661 Method calls are resolved to methods on specific traits,
2662 either statically dispatching to a method if the exact `self`-type of the left-hand-side is known,
2663 or dynamically dispatching if the left-hand-side expression is an indirect [object type](#object-types).
2665 ### Field expressions
2668 field_expr : expr '.' ident ;
2671 A _field expression_ consists of an expression followed by a single dot and an identifier,
2672 when not immediately followed by a parenthesized expression-list (the latter is a [method call expression](#method-call-expressions)).
2673 A field expression denotes a field of a [structure](#structure-types).
2675 ~~~~ {.ignore .field}
2678 (Struct {a: 10, b: 20}).a;
2681 A field access is an [lvalue](#lvalues-rvalues-and-temporaries) referring to the value of that field.
2682 When the type providing the field inherits mutabilty, it can be [assigned](#assignment-expressions) to.
2684 Also, if the type of the expression to the left of the dot is a pointer,
2685 it is automatically dereferenced to make the field access possible.
2687 ### Vector expressions
2690 vec_expr : '[' "mut" ? vec_elems? ']' ;
2692 vec_elems : [expr [',' expr]*] | [expr ',' ".." expr] ;
2695 A [_vector_](#vector-types) _expression_ is written by enclosing zero or
2696 more comma-separated expressions of uniform type in square brackets.
2698 In the `[expr ',' ".." expr]` form, the expression after the `".."`
2699 must be a constant expression that can be evaluated at compile time, such
2700 as a [literal](#literals) or a [static item](#static-items).
2704 ["a", "b", "c", "d"];
2705 [0, ..128]; // vector with 128 zeros
2706 [0u8, 0u8, 0u8, 0u8];
2709 ### Index expressions
2712 idx_expr : expr '[' expr ']' ;
2715 [Vector](#vector-types)-typed expressions can be indexed by writing a
2716 square-bracket-enclosed expression (the index) after them. When the
2717 vector is mutable, the resulting [lvalue](#lvalues-rvalues-and-temporaries) can be assigned to.
2719 Indices are zero-based, and may be of any integral type. Vector access
2720 is bounds-checked at run-time. When the check fails, it will put the
2721 task in a _failing state_.
2725 # task::spawn(proc() {
2728 (["a", "b"])[10]; // fails
2733 ### Unary operator expressions
2735 Rust defines six symbolic unary operators.
2736 They are all written as prefix operators,
2737 before the expression they apply to.
2740 : Negation. May only be applied to numeric types.
2742 : Dereference. When applied to a [pointer](#pointer-types) it denotes the pointed-to location.
2743 For pointers to mutable locations, the resulting [lvalue](#lvalues-rvalues-and-temporaries) can be assigned to.
2744 On non-pointer types, it calls the `deref` method of the `std::ops::Deref` trait, or the
2745 `deref_mut` method of the `std::ops::DerefMut` trait (if implemented by the type and required
2746 for an outer expression that will or could mutate the dereference), and produces the
2747 result of dereferencing the `&` or `&mut` borrowed pointer returned from the overload method.
2750 : Logical negation. On the boolean type, this flips between `true` and
2751 `false`. On integer types, this inverts the individual bits in the
2752 two's complement representation of the value.
2754 : [Boxing](#pointer-types) operators. Allocate a box to hold the value they are applied to,
2755 and store the value in it. `box` creates an owned box.
2757 : Borrow operator. Returns a reference, pointing to its operand.
2758 The operand of a borrow is statically proven to outlive the resulting pointer.
2759 If the borrow-checker cannot prove this, it is a compilation error.
2761 ### Binary operator expressions
2764 binop_expr : expr binop expr ;
2767 Binary operators expressions are given in terms of
2768 [operator precedence](#operator-precedence).
2770 #### Arithmetic operators
2772 Binary arithmetic expressions are syntactic sugar for calls to built-in traits,
2773 defined in the `std::ops` module of the `std` library.
2774 This means that arithmetic operators can be overridden for user-defined types.
2775 The default meaning of the operators on standard types is given here.
2778 : Addition and vector/string concatenation.
2779 Calls the `add` method on the `std::ops::Add` trait.
2782 Calls the `sub` method on the `std::ops::Sub` trait.
2785 Calls the `mul` method on the `std::ops::Mul` trait.
2788 Calls the `div` method on the `std::ops::Div` trait.
2791 Calls the `rem` method on the `std::ops::Rem` trait.
2793 #### Bitwise operators
2795 Like the [arithmetic operators](#arithmetic-operators), bitwise operators
2796 are syntactic sugar for calls to methods of built-in traits.
2797 This means that bitwise operators can be overridden for user-defined types.
2798 The default meaning of the operators on standard types is given here.
2802 Calls the `bitand` method of the `std::ops::BitAnd` trait.
2805 Calls the `bitor` method of the `std::ops::BitOr` trait.
2808 Calls the `bitxor` method of the `std::ops::BitXor` trait.
2810 : Logical left shift.
2811 Calls the `shl` method of the `std::ops::Shl` trait.
2813 : Logical right shift.
2814 Calls the `shr` method of the `std::ops::Shr` trait.
2816 #### Lazy boolean operators
2818 The operators `||` and `&&` may be applied to operands of boolean type.
2819 The `||` operator denotes logical 'or', and the `&&` operator denotes logical 'and'.
2820 They differ from `|` and `&` in that the right-hand operand is only evaluated
2821 when the left-hand operand does not already determine the result of the expression.
2822 That is, `||` only evaluates its right-hand operand
2823 when the left-hand operand evaluates to `false`, and `&&` only when it evaluates to `true`.
2825 #### Comparison operators
2827 Comparison operators are, like the [arithmetic operators](#arithmetic-operators),
2828 and [bitwise operators](#bitwise-operators),
2829 syntactic sugar for calls to built-in traits.
2830 This means that comparison operators can be overridden for user-defined types.
2831 The default meaning of the operators on standard types is given here.
2835 Calls the `eq` method on the `std::cmp::PartialEq` trait.
2838 Calls the `ne` method on the `std::cmp::PartialEq` trait.
2841 Calls the `lt` method on the `std::cmp::PartialOrd` trait.
2844 Calls the `gt` method on the `std::cmp::PartialOrd` trait.
2846 : Less than or equal.
2847 Calls the `le` method on the `std::cmp::PartialOrd` trait.
2849 : Greater than or equal.
2850 Calls the `ge` method on the `std::cmp::PartialOrd` trait.
2852 #### Type cast expressions
2854 A type cast expression is denoted with the binary operator `as`.
2856 Executing an `as` expression casts the value on the left-hand side to the type
2857 on the right-hand side.
2859 A numeric value can be cast to any numeric type.
2860 A raw pointer value can be cast to or from any integral type or raw pointer type.
2861 Any other cast is unsupported and will fail to compile.
2863 An example of an `as` expression:
2866 # fn sum(v: &[f64]) -> f64 { 0.0 }
2867 # fn len(v: &[f64]) -> int { 0 }
2869 fn avg(v: &[f64]) -> f64 {
2870 let sum: f64 = sum(v);
2871 let sz: f64 = len(v) as f64;
2876 #### Assignment expressions
2878 An _assignment expression_ consists of an [lvalue](#lvalues-rvalues-and-temporaries) expression followed by an
2879 equals sign (`=`) and an [rvalue](#lvalues-rvalues-and-temporaries) expression.
2881 Evaluating an assignment expression [either copies or moves](#moved-and-copied-types) its right-hand operand to its left-hand operand.
2890 #### Compound assignment expressions
2892 The `+`, `-`, `*`, `/`, `%`, `&`, `|`, `^`, `<<`, and `>>`
2893 operators may be composed with the `=` operator. The expression `lval
2894 OP= val` is equivalent to `lval = lval OP val`. For example, `x = x +
2895 1` may be written as `x += 1`.
2897 Any such expression always has the [`unit`](#primitive-types) type.
2899 #### Operator precedence
2901 The precedence of Rust binary operators is ordered as follows, going
2902 from strong to weak:
2904 ~~~~ {.text .precedence}
2919 Operators at the same precedence level are evaluated left-to-right. [Unary operators](#unary-operator-expressions)
2920 have the same precedence level and it is stronger than any of the binary operators'.
2922 ### Grouped expressions
2924 An expression enclosed in parentheses evaluates to the result of the enclosed
2925 expression. Parentheses can be used to explicitly specify evaluation order
2926 within an expression.
2929 paren_expr : '(' expr ')' ;
2932 An example of a parenthesized expression:
2935 let x = (2 + 3) * 4;
2939 ### Call expressions
2942 expr_list : [ expr [ ',' expr ]* ] ? ;
2943 paren_expr_list : '(' expr_list ')' ;
2944 call_expr : expr paren_expr_list ;
2947 A _call expression_ invokes a function, providing zero or more input slots and
2948 an optional reference slot to serve as the function's output, bound to the
2949 `lval` on the right hand side of the call. If the function eventually returns,
2950 then the expression completes.
2952 Some examples of call expressions:
2955 # use std::from_str::FromStr;
2956 # fn add(x: int, y: int) -> int { 0 }
2958 let x: int = add(1, 2);
2959 let pi: Option<f32> = FromStr::from_str("3.14");
2962 ### Lambda expressions
2965 ident_list : [ ident [ ',' ident ]* ] ? ;
2966 lambda_expr : '|' ident_list '|' expr ;
2969 A _lambda expression_ (sometimes called an "anonymous function expression") defines a function and denotes it as a value,
2970 in a single expression.
2971 A lambda expression is a pipe-symbol-delimited (`|`) list of identifiers followed by an expression.
2973 A lambda expression denotes a function that maps a list of parameters (`ident_list`)
2974 onto the expression that follows the `ident_list`.
2975 The identifiers in the `ident_list` are the parameters to the function.
2976 These parameters' types need not be specified, as the compiler infers them from context.
2978 Lambda expressions are most useful when passing functions as arguments to other functions,
2979 as an abbreviation for defining and capturing a separate function.
2981 Significantly, lambda expressions _capture their environment_,
2982 which regular [function definitions](#functions) do not.
2983 The exact type of capture depends on the [function type](#function-types) inferred for the lambda expression.
2984 In the simplest and least-expensive form (analogous to a ```|| { }``` expression),
2985 the lambda expression captures its environment by reference,
2986 effectively borrowing pointers to all outer variables mentioned inside the function.
2987 Alternately, the compiler may infer that a lambda expression should copy or move values (depending on their type.)
2988 from the environment into the lambda expression's captured environment.
2990 In this example, we define a function `ten_times` that takes a higher-order function argument,
2991 and call it with a lambda expression as an argument.
2994 fn ten_times(f: |int|) {
3002 ten_times(|j| println!("hello, {}", j));
3008 while_expr : "while" no_struct_literal_expr '{' block '}' ;
3011 A `while` loop begins by evaluating the boolean loop conditional expression.
3012 If the loop conditional expression evaluates to `true`, the loop body block
3013 executes and control returns to the loop conditional expression. If the loop
3014 conditional expression evaluates to `false`, the `while` expression completes.
3029 A `loop` expression denotes an infinite loop.
3032 loop_expr : [ lifetime ':' ] "loop" '{' block '}';
3035 A `loop` expression may optionally have a _label_.
3036 If a label is present,
3037 then labeled `break` and `continue` expressions nested within this loop may exit out of this loop or return control to its head.
3038 See [Break expressions](#break-expressions) and [Continue expressions](#continue-expressions).
3040 ### Break expressions
3043 break_expr : "break" [ lifetime ];
3046 A `break` expression has an optional _label_.
3047 If the label is absent, then executing a `break` expression immediately terminates the innermost loop enclosing it.
3048 It is only permitted in the body of a loop.
3049 If the label is present, then `break foo` terminates the loop with label `foo`,
3050 which need not be the innermost label enclosing the `break` expression,
3051 but must enclose it.
3053 ### Continue expressions
3056 continue_expr : "continue" [ lifetime ];
3059 A `continue` expression has an optional _label_.
3060 If the label is absent,
3061 then executing a `continue` expression immediately terminates the current iteration of the innermost loop enclosing it,
3062 returning control to the loop *head*.
3063 In the case of a `while` loop,
3064 the head is the conditional expression controlling the loop.
3065 In the case of a `for` loop, the head is the call-expression controlling the loop.
3066 If the label is present, then `continue foo` returns control to the head of the loop with label `foo`,
3067 which need not be the innermost label enclosing the `break` expression,
3068 but must enclose it.
3070 A `continue` expression is only permitted in the body of a loop.
3075 for_expr : "for" pat "in" no_struct_literal_expr '{' block '}' ;
3078 A `for` expression is a syntactic construct for looping over elements
3079 provided by an implementation of `std::iter::Iterator`.
3081 An example of a for loop over the contents of a vector:
3085 # fn bar(f: Foo) { }
3090 let v: &[Foo] = &[a, b, c];
3097 An example of a for loop over a series of integers:
3100 # fn bar(b:uint) { }
3101 for i in range(0u, 256) {
3109 if_expr : "if" no_struct_literal_expr '{' block '}'
3112 else_tail : "else" [ if_expr
3116 An `if` expression is a conditional branch in program control. The form of
3117 an `if` expression is a condition expression, followed by a consequent
3118 block, any number of `else if` conditions and blocks, and an optional
3119 trailing `else` block. The condition expressions must have type
3120 `bool`. If a condition expression evaluates to `true`, the
3121 consequent block is executed and any subsequent `else if` or `else`
3122 block is skipped. If a condition expression evaluates to `false`, the
3123 consequent block is skipped and any subsequent `else if` condition is
3124 evaluated. If all `if` and `else if` conditions evaluate to `false`
3125 then any `else` block is executed.
3127 ### Match expressions
3130 match_expr : "match" no_struct_literal_expr '{' match_arm * '}' ;
3132 match_arm : attribute * match_pat "=>" [ expr "," | '{' block '}' ] ;
3134 match_pat : pat [ '|' pat ] * [ "if" expr ] ? ;
3137 A `match` expression branches on a *pattern*. The exact form of matching that
3138 occurs depends on the pattern. Patterns consist of some combination of
3139 literals, destructured vectors or enum constructors, structures and
3140 tuples, variable binding specifications, wildcards (`..`), and placeholders
3141 (`_`). A `match` expression has a *head expression*, which is the value to
3142 compare to the patterns. The type of the patterns must equal the type of the
3145 In a pattern whose head expression has an `enum` type, a placeholder (`_`)
3146 stands for a *single* data field, whereas a wildcard `..` stands for *all* the
3147 fields of a particular variant. For example:
3150 enum List<X> { Nil, Cons(X, Box<List<X>>) }
3152 let x: List<int> = Cons(10, box Cons(11, box Nil));
3155 Cons(_, box Nil) => fail!("singleton list"),
3157 Nil => fail!("empty list")
3161 The first pattern matches lists constructed by applying `Cons` to any head
3162 value, and a tail value of `box Nil`. The second pattern matches _any_ list
3163 constructed with `Cons`, ignoring the values of its arguments. The difference
3164 between `_` and `..` is that the pattern `C(_)` is only type-correct if `C` has
3165 exactly one argument, while the pattern `C(..)` is type-correct for any enum
3166 variant `C`, regardless of how many arguments `C` has.
3168 Used inside a vector pattern, `..` stands for any number of elements. This
3169 wildcard can be used at most once for a given vector, which implies that it
3170 cannot be used to specifically match elements that are at an unknown distance
3171 from both ends of a vector, like `[.., 42, ..]`. If followed by a variable name,
3172 it will bind the corresponding slice to the variable. Example:
3175 fn is_symmetric(list: &[uint]) -> bool {
3178 [x, ..inside, y] if x == y => is_symmetric(inside),
3184 let sym = &[0, 1, 4, 2, 4, 1, 0];
3185 let not_sym = &[0, 1, 7, 2, 4, 1, 0];
3186 assert!(is_symmetric(sym));
3187 assert!(!is_symmetric(not_sym));
3191 A `match` behaves differently depending on whether or not the head expression
3192 is an [lvalue or an rvalue](#lvalues-rvalues-and-temporaries).
3193 If the head expression is an rvalue, it is
3194 first evaluated into a temporary location, and the resulting value
3195 is sequentially compared to the patterns in the arms until a match
3196 is found. The first arm with a matching pattern is chosen as the branch target
3197 of the `match`, any variables bound by the pattern are assigned to local
3198 variables in the arm's block, and control enters the block.
3200 When the head expression is an lvalue, the match does not allocate a
3201 temporary location (however, a by-value binding may copy or move from
3202 the lvalue). When possible, it is preferable to match on lvalues, as the
3203 lifetime of these matches inherits the lifetime of the lvalue, rather
3204 than being restricted to the inside of the match.
3206 An example of a `match` expression:
3209 # fn process_pair(a: int, b: int) { }
3210 # fn process_ten() { }
3212 enum List<X> { Nil, Cons(X, Box<List<X>>) }
3214 let x: List<int> = Cons(10, box Cons(11, box Nil));
3217 Cons(a, box Cons(b, _)) => {
3232 Patterns that bind variables
3233 default to binding to a copy or move of the matched value
3234 (depending on the matched value's type).
3235 This can be changed to bind to a reference by
3236 using the `ref` keyword,
3237 or to a mutable reference using `ref mut`.
3239 Subpatterns can also be bound to variables by the use of the syntax
3240 `variable @ subpattern`.
3244 enum List { Nil, Cons(uint, Box<List>) }
3246 fn is_sorted(list: &List) -> bool {
3248 Nil | Cons(_, box Nil) => true,
3249 Cons(x, ref r @ box Cons(y, _)) => (x <= y) && is_sorted(*r)
3254 let a = Cons(6, box Cons(7, box Cons(42, box Nil)));
3255 assert!(is_sorted(&a));
3260 Patterns can also dereference pointers by using the `&`,
3261 `box` or `@` symbols, as appropriate. For example, these two matches
3262 on `x: &int` are equivalent:
3266 let y = match *x { 0 => "zero", _ => "some" };
3267 let z = match x { &0 => "zero", _ => "some" };
3272 A pattern that's just an identifier, like `Nil` in the previous example,
3273 could either refer to an enum variant that's in scope, or bind a new variable.
3274 The compiler resolves this ambiguity by forbidding variable bindings that occur
3275 in `match` patterns from shadowing names of variants that are in scope.
3276 For example, wherever `List` is in scope,
3277 a `match` pattern would not be able to bind `Nil` as a new name.
3278 The compiler interprets a variable pattern `x` as a binding _only_ if there is
3279 no variant named `x` in scope.
3280 A convention you can use to avoid conflicts is simply to name variants with
3281 upper-case letters, and local variables with lower-case letters.
3283 Multiple match patterns may be joined with the `|` operator.
3284 A range of values may be specified with `..`.
3290 let message = match x {
3291 0 | 1 => "not many",
3297 Range patterns only work on scalar types
3298 (like integers and characters; not like vectors and structs, which have sub-components).
3299 A range pattern may not be a sub-range of another range pattern inside the same `match`.
3301 Finally, match patterns can accept *pattern guards* to further refine the
3302 criteria for matching a case. Pattern guards appear after the pattern and
3303 consist of a bool-typed expression following the `if` keyword. A pattern
3304 guard may refer to the variables bound within the pattern they follow.
3307 # let maybe_digit = Some(0);
3308 # fn process_digit(i: int) { }
3309 # fn process_other(i: int) { }
3311 let message = match maybe_digit {
3312 Some(x) if x < 10 => process_digit(x),
3313 Some(x) => process_other(x),
3318 ### Return expressions
3321 return_expr : "return" expr ? ;
3324 Return expressions are denoted with the keyword `return`. Evaluating a `return`
3325 expression moves its argument into the output slot of the current
3326 function, destroys the current function activation frame, and transfers
3327 control to the caller frame.
3329 An example of a `return` expression:
3332 fn max(a: int, b: int) -> int {
3344 Every slot, item and value in a Rust program has a type. The _type_ of a *value*
3345 defines the interpretation of the memory holding it.
3347 Built-in types and type-constructors are tightly integrated into the language,
3348 in nontrivial ways that are not possible to emulate in user-defined
3349 types. User-defined types have limited capabilities.
3353 The primitive types are the following:
3355 * The "unit" type `()`, having the single "unit" value `()` (occasionally called
3357 * The boolean type `bool` with values `true` and `false`.
3358 * The machine types.
3359 * The machine-dependent integer and floating-point types.
3361 [^unittype]: The "unit" value `()` is *not* a sentinel "null pointer" value for
3362 reference slots; the "unit" type is the implicit return type from functions
3363 otherwise lacking a return type, and can be used in other contexts (such as
3364 message-sending or type-parametric code) as a zero-size type.]
3368 The machine types are the following:
3370 * The unsigned word types `u8`, `u16`, `u32` and `u64`, with values drawn from
3371 the integer intervals [0, 2^8 - 1], [0, 2^16 - 1], [0, 2^32 - 1] and
3372 [0, 2^64 - 1] respectively.
3374 * The signed two's complement word types `i8`, `i16`, `i32` and `i64`, with
3375 values drawn from the integer intervals [-(2^(7)), 2^7 - 1],
3376 [-(2^(15)), 2^15 - 1], [-(2^(31)), 2^31 - 1], [-(2^(63)), 2^63 - 1]
3379 * The IEEE 754-2008 `binary32` and `binary64` floating-point types: `f32` and
3380 `f64`, respectively.
3382 #### Machine-dependent integer types
3384 The Rust type `uint` [^rustuint] is an
3385 unsigned integer type with target-machine-dependent size. Its size, in
3386 bits, is equal to the number of bits required to hold any memory address on
3389 The Rust type `int` [^rustint] is a
3390 two's complement signed integer type with target-machine-dependent size. Its
3391 size, in bits, is equal to the size of the rust type `uint` on the same target
3394 [^rustuint]: A Rust `uint` is analogous to a C99 `uintptr_t`.
3395 [^rustint]: A Rust `int` is analogous to a C99 `intptr_t`.
3399 The types `char` and `str` hold textual data.
3401 A value of type `char` is a [Unicode scalar value](
3402 http://www.unicode.org/glossary/#unicode_scalar_value)
3403 (ie. a code point that is not a surrogate),
3404 represented as a 32-bit unsigned word in the 0x0000 to 0xD7FF
3405 or 0xE000 to 0x10FFFF range.
3406 A `[char]` vector is effectively an UCS-4 / UTF-32 string.
3408 A value of type `str` is a Unicode string,
3409 represented as a vector of 8-bit unsigned bytes holding a sequence of UTF-8 codepoints.
3410 Since `str` is of unknown size, it is not a _first class_ type,
3411 but can only be instantiated through a pointer type,
3412 such as `&str` or `String`.
3416 A tuple *type* is a heterogeneous product of other types, called the *elements*
3417 of the tuple. It has no nominal name and is instead structurally typed.
3419 Tuple types and values are denoted by listing the types or values of their
3420 elements, respectively, in a parenthesized, comma-separated
3423 Because tuple elements don't have a name, they can only be accessed by pattern-matching.
3425 The members of a tuple are laid out in memory contiguously, in
3426 order specified by the tuple type.
3428 An example of a tuple type and its use:
3431 type Pair<'a> = (int, &'a str);
3432 let p: Pair<'static> = (10, "hello");
3434 assert!(b != "world");
3439 The vector type constructor represents a homogeneous array of values of a given type.
3440 A vector has a fixed size.
3441 (Operations like `vec.push` operate solely on owned vectors.)
3442 A vector type can be annotated with a _definite_ size, such as `[int, ..10]`.
3443 Such a definite-sized vector type is a first-class type, since its size is known statically.
3444 A vector without such a size is said to be of _indefinite_ size,
3445 and is therefore not a _first-class_ type.
3446 An indefinite-size vector can only be instantiated through a pointer type,
3447 such as `&[T]` or `Vec<T>`.
3448 The kind of a vector type depends on the kind of its element type,
3449 as with other simple structural types.
3451 Expressions producing vectors of definite size cannot be evaluated in a
3452 context expecting a vector of indefinite size; one must copy the
3453 definite-sized vector contents into a distinct vector of indefinite size.
3455 An example of a vector type and its use:
3458 let v: &[int] = &[7, 5, 3];
3463 All in-bounds elements of a vector are always initialized,
3464 and access to a vector is always bounds-checked.
3468 A `struct` *type* is a heterogeneous product of other types, called the *fields*
3469 of the type.[^structtype]
3471 [^structtype]: `struct` types are analogous `struct` types in C,
3472 the *record* types of the ML family,
3473 or the *structure* types of the Lisp family.
3475 New instances of a `struct` can be constructed with a [struct expression](#structure-expressions).
3477 The memory layout of a `struct` is undefined by default to allow for compiler optimziations like
3478 field reordering, but it can be fixed with the `#[repr(...)]` attribute.
3479 In either case, fields may be given in any order in a corresponding struct *expression*;
3480 the resulting `struct` value will always have the same memory layout.
3482 The fields of a `struct` may be qualified by [visibility modifiers](#re-exporting-and-visibility),
3483 to allow access to data in a structure outside a module.
3485 A _tuple struct_ type is just like a structure type, except that the fields are anonymous.
3487 A _unit-like struct_ type is like a structure type, except that it has no fields.
3488 The one value constructed by the associated [structure expression](#structure-expressions)
3489 is the only value that inhabits such a type.
3491 ### Enumerated types
3493 An *enumerated type* is a nominal, heterogeneous disjoint union type,
3494 denoted by the name of an [`enum` item](#enumerations). [^enumtype]
3496 [^enumtype]: The `enum` type is analogous to a `data` constructor declaration in
3497 ML, or a *pick ADT* in Limbo.
3499 An [`enum` item](#enumerations) declares both the type and a number of *variant constructors*,
3500 each of which is independently named and takes an optional tuple of arguments.
3502 New instances of an `enum` can be constructed by calling one of the variant constructors,
3503 in a [call expression](#call-expressions).
3505 Any `enum` value consumes as much memory as the largest variant constructor for its corresponding `enum` type.
3507 Enum types cannot be denoted *structurally* as types,
3508 but must be denoted by named reference to an [`enum` item](#enumerations).
3512 Nominal types — [enumerations](#enumerated-types) and [structures](#structure-types) — may be recursive.
3513 That is, each `enum` constructor or `struct` field may refer, directly or indirectly, to the enclosing `enum` or `struct` type itself.
3514 Such recursion has restrictions:
3516 * Recursive types must include a nominal type in the recursion
3517 (not mere [type definitions](#type-definitions),
3518 or other structural types such as [vectors](#vector-types) or [tuples](#tuple-types)).
3519 * A recursive `enum` item must have at least one non-recursive constructor
3520 (in order to give the recursion a basis case).
3521 * The size of a recursive type must be finite;
3522 in other words the recursive fields of the type must be [pointer types](#pointer-types).
3523 * Recursive type definitions can cross module boundaries, but not module *visibility* boundaries,
3524 or crate boundaries (in order to simplify the module system and type checker).
3526 An example of a *recursive* type and its use:
3531 Cons(T, Box<List<T>>)
3534 let a: List<int> = Cons(7, box Cons(13, box Nil));
3539 All pointers in Rust are explicit first-class values.
3540 They can be copied, stored into data structures, and returned from functions.
3541 There are four varieties of pointer in Rust:
3543 * Owning pointers (`Box`)
3544 : These point to owned heap allocations (or "boxes") in the shared, inter-task heap.
3545 Each owned box has a single owning pointer; pointer and pointee retain a 1:1 relationship at all times.
3546 Owning pointers are written `Box<content>`,
3547 for example `Box<int>` means an owning pointer to an owned box containing an integer.
3548 Copying an owned box is a "deep" operation:
3549 it involves allocating a new owned box and copying the contents of the old box into the new box.
3550 Releasing an owning pointer immediately releases its corresponding owned box.
3553 : These point to memory _owned by some other value_.
3554 References arise by (automatic) conversion from owning pointers, managed pointers,
3555 or by applying the borrowing operator `&` to some other value,
3556 including [lvalues, rvalues or temporaries](#lvalues-rvalues-and-temporaries).
3557 References are written `&content`, or in some cases `&'f content` for some lifetime-variable `f`,
3558 for example `&int` means a reference to an integer.
3559 Copying a reference is a "shallow" operation:
3560 it involves only copying the pointer itself.
3561 Releasing a reference typically has no effect on the value it points to,
3562 with the exception of temporary values,
3563 which are released when the last reference to them is released.
3565 * Raw pointers (`*`)
3566 : Raw pointers are pointers without safety or liveness guarantees.
3567 Raw pointers are written as `*const T` or `*mut T`,
3568 for example `*const int` means a raw pointer to an integer.
3569 Copying or dropping a raw pointer has no effect on the lifecycle of any
3570 other value. Dereferencing a raw pointer or converting it to any other
3571 pointer type is an [`unsafe` operation](#unsafe-functions).
3572 Raw pointers are generally discouraged in Rust code;
3573 they exist to support interoperability with foreign code,
3574 and writing performance-critical or low-level functions.
3578 The function type constructor `fn` forms new function types.
3579 A function type consists of a possibly-empty set of function-type modifiers
3580 (such as `unsafe` or `extern`), a sequence of input types and an output type.
3582 An example of a `fn` type:
3585 fn add(x: int, y: int) -> int {
3589 let mut x = add(5,7);
3591 type Binop<'a> = |int,int|: 'a -> int;
3592 let bo: Binop = add;
3598 ~~~~ {.ebnf .notation}
3599 closure_type := [ 'unsafe' ] [ '<' lifetime-list '>' ] '|' arg-list '|'
3600 [ ':' bound-list ] [ '->' type ]
3601 procedure_type := 'proc' [ '<' lifetime-list '>' ] '(' arg-list ')'
3602 [ ':' bound-list ] [ '->' type ]
3603 lifetime-list := lifetime | lifetime ',' lifetime-list
3604 arg-list := ident ':' type | ident ':' type ',' arg-list
3605 bound-list := bound | bound '+' bound-list
3606 bound := path | lifetime
3609 The type of a closure mapping an input of type `A` to an output of type `B` is
3610 `|A| -> B`. A closure with no arguments or return values has type `||`.
3611 Similarly, a procedure mapping `A` to `B` is `proc(A) -> B` and a no-argument
3612 and no-return value closure has type `proc()`.
3614 An example of creating and calling a closure:
3617 let captured_var = 10i;
3619 let closure_no_args = || println!("captured_var={}", captured_var);
3621 let closure_args = |arg: int| -> int {
3622 println!("captured_var={}, arg={}", captured_var, arg);
3623 arg // Note lack of semicolon after 'arg'
3626 fn call_closure(c1: ||, c2: |int| -> int) {
3631 call_closure(closure_no_args, closure_args);
3635 Unlike closures, procedures may only be invoked once, but own their
3636 environment, and are allowed to move out of their environment. Procedures are
3637 allocated on the heap (unlike closures). An example of creating and calling a
3641 let string = "Hello".to_string();
3643 // Creates a new procedure, passing it to the `spawn` function.
3645 println!("{} world!", string);
3648 // the variable `string` has been moved into the previous procedure, so it is
3649 // no longer usable.
3652 // Create an invoke a procedure. Note that the procedure is *moved* when
3653 // invoked, so it cannot be invoked again.
3654 let f = proc(n: int) { n + 22 };
3655 println!("answer: {}", f(20));
3661 Every trait item (see [traits](#traits)) defines a type with the same name as the trait.
3662 This type is called the _object type_ of the trait.
3663 Object types permit "late binding" of methods, dispatched using _virtual method tables_ ("vtables").
3664 Whereas most calls to trait methods are "early bound" (statically resolved) to specific implementations at compile time,
3665 a call to a method on an object type is only resolved to a vtable entry at compile time.
3666 The actual implementation for each vtable entry can vary on an object-by-object basis.
3668 Given a pointer-typed expression `E` of type `&T` or `Box<T>`, where `T` implements trait `R`,
3669 casting `E` to the corresponding pointer type `&R` or `Box<R>` results in a value of the _object type_ `R`.
3670 This result is represented as a pair of pointers:
3671 the vtable pointer for the `T` implementation of `R`, and the pointer value of `E`.
3673 An example of an object type:
3677 fn to_string(&self) -> String;
3680 impl Printable for int {
3681 fn to_string(&self) -> String { self.to_str() }
3684 fn print(a: Box<Printable>) {
3685 println!("{}", a.to_string());
3689 print(box 10i as Box<Printable>);
3693 In this example, the trait `Printable` occurs as an object type in both the type signature of `print`,
3694 and the cast expression in `main`.
3698 Within the body of an item that has type parameter declarations, the names of its type parameters are types:
3701 fn map<A: Clone, B: Clone>(f: |A| -> B, xs: &[A]) -> Vec<B> {
3705 let first: B = f(xs[0].clone());
3706 let rest: Vec<B> = map(f, xs.slice(1, xs.len()));
3707 return vec![first].append(rest.as_slice());
3711 Here, `first` has type `B`, referring to `map`'s `B` type parameter;
3712 and `rest` has type `Vec<B>`, a vector type with element type `B`.
3716 The special type `self` has a meaning within methods inside an
3717 impl item. It refers to the type of the implicit `self` argument. For
3722 fn make_string(&self) -> String;
3725 impl Printable for String {
3726 fn make_string(&self) -> String {
3732 `self` refers to the value of type `String` that is the receiver for a
3733 call to the method `make_string`.
3737 Types in Rust are categorized into kinds, based on various properties of the components of the type.
3741 : Types of this kind can be safely sent between tasks.
3742 This kind includes scalars, owning pointers, owned closures, and
3743 structural types containing only other owned types.
3744 All `Send` types are `'static`.
3746 : Types of this kind consist of "Plain Old Data"
3747 which can be copied by simply moving bits.
3748 All values of this kind can be implicitly copied.
3749 This kind includes scalars and immutable references,
3750 as well as structural types containing other `Copy` types.
3752 : Types of this kind do not contain any references (except for
3753 references with the `static` lifetime, which are allowed).
3754 This can be a useful guarantee for code
3755 that breaks borrowing assumptions
3756 using [`unsafe` operations](#unsafe-functions).
3758 : This is not strictly a kind,
3759 but its presence interacts with kinds:
3760 the `Drop` trait provides a single method `drop`
3761 that takes no parameters,
3762 and is run when values of the type are dropped.
3763 Such a method is called a "destructor",
3764 and are always executed in "top-down" order:
3765 a value is completely destroyed
3766 before any of the values it owns run their destructors.
3767 Only `Send` types can implement `Drop`.
3770 : Types with destructors, closure environments,
3771 and various other _non-first-class_ types,
3772 are not copyable at all.
3773 Such types can usually only be accessed through pointers,
3774 or in some cases, moved between mutable locations.
3776 Kinds can be supplied as _bounds_ on type parameters, like traits,
3777 in which case the parameter is constrained to types satisfying that kind.
3779 By default, type parameters do not carry any assumed kind-bounds at all.
3780 When instantiating a type parameter,
3781 the kind bounds on the parameter are checked
3782 to be the same or narrower than the kind
3783 of the type that it is instantiated with.
3785 Sending operations are not part of the Rust language,
3786 but are implemented in the library.
3787 Generic functions that send values
3788 bound the kind of these values to sendable.
3790 # Memory and concurrency models
3792 Rust has a memory model centered around concurrently-executing _tasks_. Thus
3793 its memory model and its concurrency model are best discussed simultaneously,
3794 as parts of each only make sense when considered from the perspective of the
3797 When reading about the memory model, keep in mind that it is partitioned in
3798 order to support tasks; and when reading about tasks, keep in mind that their
3799 isolation and communication mechanisms are only possible due to the ownership
3800 and lifetime semantics of the memory model.
3804 A Rust program's memory consists of a static set of *items*, a set of
3805 [tasks](#tasks) each with its own *stack*, and a *heap*. Immutable portions of
3806 the heap may be shared between tasks, mutable portions may not.
3808 Allocations in the stack consist of *slots*, and allocations in the heap
3811 ### Memory allocation and lifetime
3813 The _items_ of a program are those functions, modules and types
3814 that have their value calculated at compile-time and stored uniquely in the
3815 memory image of the rust process. Items are neither dynamically allocated nor
3818 A task's _stack_ consists of activation frames automatically allocated on
3819 entry to each function as the task executes. A stack allocation is reclaimed
3820 when control leaves the frame containing it.
3822 The _heap_ is a general term that describes two separate sets of boxes:
3823 managed boxes — which may be subject to garbage collection — and owned
3824 boxes. The lifetime of an allocation in the heap depends on the lifetime of
3825 the box values pointing to it. Since box values may themselves be passed in
3826 and out of frames, or stored in the heap, heap allocations may outlive the
3827 frame they are allocated within.
3829 ### Memory ownership
3831 A task owns all memory it can *safely* reach through local variables,
3832 as well as managed, owned boxes and references.
3834 When a task sends a value that has the `Send` trait to another task,
3835 it loses ownership of the value sent and can no longer refer to it.
3836 This is statically guaranteed by the combined use of "move semantics",
3837 and the compiler-checked _meaning_ of the `Send` trait:
3838 it is only instantiated for (transitively) sendable kinds of data constructor and pointers,
3839 never including managed boxes or references.
3841 When a stack frame is exited, its local allocations are all released, and its
3842 references to boxes (both managed and owned) are dropped.
3844 A managed box may (in the case of a recursive, mutable managed type) be cyclic;
3845 in this case the release of memory inside the managed structure may be deferred
3846 until task-local garbage collection can reclaim it. Code can ensure no such
3847 delayed deallocation occurs by restricting itself to owned boxes and similar
3848 unmanaged kinds of data.
3850 When a task finishes, its stack is necessarily empty and it therefore has no
3851 references to any boxes; the remainder of its heap is immediately freed.
3855 A task's stack contains slots.
3857 A _slot_ is a component of a stack frame, either a function parameter,
3858 a [temporary](#lvalues-rvalues-and-temporaries), or a local variable.
3860 A _local variable_ (or *stack-local* allocation) holds a value directly,
3861 allocated within the stack's memory. The value is a part of the stack frame.
3863 Local variables are immutable unless declared otherwise like: `let mut x = ...`.
3865 Function parameters are immutable unless declared with `mut`. The
3866 `mut` keyword applies only to the following parameter (so `|mut x, y|`
3867 and `fn f(mut x: Box<int>, y: Box<int>)` declare one mutable variable `x` and
3868 one immutable variable `y`).
3870 Methods that take either `self` or `~self` can optionally place them in a
3871 mutable slot by prefixing them with `mut` (similar to regular arguments):
3875 fn change(mut self) -> Self;
3876 fn modify(mut ~self) -> Box<Self>;
3880 Local variables are not initialized when allocated; the entire frame worth of
3881 local variables are allocated at once, on frame-entry, in an uninitialized
3882 state. Subsequent statements within a function may or may not initialize the
3883 local variables. Local variables can be used only after they have been
3884 initialized; this is enforced by the compiler.
3888 An _owned box_ is a reference to a heap allocation holding another value, which is constructed
3889 by the prefix operator `box`. When the standard library is in use, the type of an owned box is
3890 `std::owned::Box<T>`.
3892 An example of an owned box type and value:
3896 let x: Box<int> = box 10;
3899 Owned box values exist in 1:1 correspondence with their heap allocation
3900 copying an owned box value makes a shallow copy of the pointer
3901 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.
3904 let x: Box<int> = box 10;
3906 // attempting to use `x` will result in an error here
3913 An executing Rust program consists of a tree of tasks.
3914 A Rust _task_ consists of an entry function, a stack,
3915 a set of outgoing communication channels and incoming communication ports,
3916 and ownership of some portion of the heap of a single operating-system process.
3917 (We expect that many programs will not use channels and ports directly,
3918 but will instead use higher-level abstractions provided in standard libraries,
3921 Multiple Rust tasks may coexist in a single operating-system process.
3922 The runtime scheduler maps tasks to a certain number of operating-system threads.
3923 By default, the scheduler chooses the number of threads based on
3924 the number of concurrent physical CPUs detected at startup.
3925 It's also possible to override this choice at runtime.
3926 When the number of tasks exceeds the number of threads — which is likely —
3927 the scheduler multiplexes the tasks onto threads.[^mnscheduler]
3929 [^mnscheduler]: This is an M:N scheduler, which is known to give suboptimal
3930 results for CPU-bound concurrency problems. In such cases, running with the
3931 same number of threads and tasks can yield better results. Rust has M:N
3932 scheduling in order to support very large numbers of tasks in contexts where
3933 threads are too resource-intensive to use in large number. The cost of
3934 threads varies substantially per operating system, and is sometimes quite
3935 low, so this flexibility is not always worth exploiting.
3937 ### Communication between tasks
3939 Rust tasks are isolated and generally unable to interfere with one another's memory directly,
3940 except through [`unsafe` code](#unsafe-functions).
3941 All contact between tasks is mediated by safe forms of ownership transfer,
3942 and data races on memory are prohibited by the type system.
3944 Inter-task communication and co-ordination facilities are provided in the standard library.
3947 - synchronous and asynchronous communication channels with various communication topologies
3948 - read-only and read-write shared variables with various safe mutual exclusion patterns
3949 - simple locks and semaphores
3951 When such facilities carry values, the values are restricted to the [`Send` type-kind](#type-kinds).
3952 Restricting communication interfaces to this kind ensures that no references or managed pointers move between tasks.
3953 Thus access to an entire data structure can be mediated through its owning "root" value;
3954 no further locking or copying is required to avoid data races within the substructure of such a value.
3958 The _lifecycle_ of a task consists of a finite set of states and events
3959 that cause transitions between the states. The lifecycle states of a task are:
3966 A task begins its lifecycle — once it has been spawned — in the *running*
3967 state. In this state it executes the statements of its entry function, and any
3968 functions called by the entry function.
3970 A task may transition from the *running* state to the *blocked*
3971 state any time it makes a blocking communication call. When the
3972 call can be completed — when a message arrives at a sender, or a
3973 buffer opens to receive a message — then the blocked task will
3974 unblock and transition back to *running*.
3976 A task may transition to the *failing* state at any time, due being
3977 killed by some external event or internally, from the evaluation of a
3978 `fail!()` macro. Once *failing*, a task unwinds its stack and
3979 transitions to the *dead* state. Unwinding the stack of a task is done by
3980 the task itself, on its own control stack. If a value with a destructor is
3981 freed during unwinding, the code for the destructor is run, also on the task's
3982 control stack. Running the destructor code causes a temporary transition to a
3983 *running* state, and allows the destructor code to cause any subsequent
3984 state transitions. The original task of unwinding and failing thereby may
3985 suspend temporarily, and may involve (recursive) unwinding of the stack of a
3986 failed destructor. Nonetheless, the outermost unwinding activity will continue
3987 until the stack is unwound and the task transitions to the *dead*
3988 state. There is no way to "recover" from task failure. Once a task has
3989 temporarily suspended its unwinding in the *failing* state, failure
3990 occurring from within this destructor results in *hard* failure.
3991 A hard failure currently results in the process aborting.
3993 A task in the *dead* state cannot transition to other states; it exists
3994 only to have its termination status inspected by other tasks, and/or to await
3995 reclamation when the last reference to it drops.
3999 The currently scheduled task is given a finite *time slice* in which to
4000 execute, after which it is *descheduled* at a loop-edge or similar
4001 preemption point, and another task within is scheduled, pseudo-randomly.
4003 An executing task can yield control at any time, by making a library call to
4004 `std::task::yield`, which deschedules it immediately. Entering any other
4005 non-executing state (blocked, dead) similarly deschedules the task.
4007 # Runtime services, linkage and debugging
4009 The Rust _runtime_ is a relatively compact collection of C++ and Rust code
4010 that provides fundamental services and datatypes to all Rust tasks at
4011 run-time. It is smaller and simpler than many modern language runtimes. It is
4012 tightly integrated into the language's execution model of memory, tasks,
4013 communication and logging.
4015 > **Note:** The runtime library will merge with the `std` library in future versions of Rust.
4017 ### Memory allocation
4019 The runtime memory-management system is based on a _service-provider interface_,
4020 through which the runtime requests blocks of memory from its environment
4021 and releases them back to its environment when they are no longer needed.
4022 The default implementation of the service-provider interface
4023 consists of the C runtime functions `malloc` and `free`.
4025 The runtime memory-management system, in turn, supplies Rust tasks with
4026 facilities for allocating releasing stacks, as well as allocating and freeing
4031 The runtime provides C and Rust code to assist with various built-in types,
4032 such as vectors, strings, and the low level communication system (ports,
4035 Support for other built-in types such as simple types, tuples and
4036 enums is open-coded by the Rust compiler.
4038 ### Task scheduling and communication
4040 The runtime provides code to manage inter-task communication. This includes
4041 the system of task-lifecycle state transitions depending on the contents of
4042 queues, as well as code to copy values between queues and their recipients and
4043 to serialize values for transmission over operating-system inter-process
4044 communication facilities.
4048 The Rust compiler supports various methods to link crates together both
4049 statically and dynamically. This section will explore the various methods to
4050 link Rust crates together, and more information about native libraries can be
4051 found in the [ffi tutorial][ffi].
4053 In one session of compilation, the compiler can generate multiple artifacts
4054 through the usage of either command line flags or the `crate_type` attribute.
4055 If one or more command line flag is specified, all `crate_type` attributes will
4056 be ignored in favor of only building the artifacts specified by command line.
4058 * `--crate-type=bin`, `#[crate_type = "bin"]` - A runnable executable will be
4059 produced. This requires that there is a `main` function in the crate which
4060 will be run when the program begins executing. This will link in all Rust and
4061 native dependencies, producing a distributable binary.
4063 * `--crate-type=lib`, `#[crate_type = "lib"]` - A Rust library will be produced.
4064 This is an ambiguous concept as to what exactly is produced because a library
4065 can manifest itself in several forms. The purpose of this generic `lib` option
4066 is to generate the "compiler recommended" style of library. The output library
4067 will always be usable by rustc, but the actual type of library may change from
4068 time-to-time. The remaining output types are all different flavors of
4069 libraries, and the `lib` type can be seen as an alias for one of them (but the
4070 actual one is compiler-defined).
4072 * `--crate-type=dylib`, `#[crate_type = "dylib"]` - A dynamic Rust library will
4073 be produced. This is different from the `lib` output type in that this forces
4074 dynamic library generation. The resulting dynamic library can be used as a
4075 dependency for other libraries and/or executables. This output type will
4076 create `*.so` files on linux, `*.dylib` files on osx, and `*.dll` files on
4079 * `--crate-type=staticlib`, `#[crate_type = "staticlib"]` - A static system
4080 library will be produced. This is different from other library outputs in that
4081 the Rust compiler will never attempt to link to `staticlib` outputs. The
4082 purpose of this output type is to create a static library containing all of
4083 the local crate's code along with all upstream dependencies. The static
4084 library is actually a `*.a` archive on linux and osx and a `*.lib` file on
4085 windows. This format is recommended for use in situtations such as linking
4086 Rust code into an existing non-Rust application because it will not have
4087 dynamic dependencies on other Rust code.
4089 * `--crate-type=rlib`, `#[crate_type = "rlib"]` - A "Rust library" file will be
4090 produced. This is used as an intermediate artifact and can be thought of as a
4091 "static Rust library". These `rlib` files, unlike `staticlib` files, are
4092 interpreted by the Rust compiler in future linkage. This essentially means
4093 that `rustc` will look for metadata in `rlib` files like it looks for metadata
4094 in dynamic libraries. This form of output is used to produce statically linked
4095 executables as well as `staticlib` outputs.
4097 Note that these outputs are stackable in the sense that if multiple are
4098 specified, then the compiler will produce each form of output at once without
4099 having to recompile. However, this only applies for outputs specified by the same
4100 method. If only `crate_type` attributes are specified, then they will all be
4101 built, but if one or more `--crate-type` command line flag is specified,
4102 then only those outputs will be built.
4104 With all these different kinds of outputs, if crate A depends on crate B, then
4105 the compiler could find B in various different forms throughout the system. The
4106 only forms looked for by the compiler, however, are the `rlib` format and the
4107 dynamic library format. With these two options for a dependent library, the
4108 compiler must at some point make a choice between these two formats. With this
4109 in mind, the compiler follows these rules when determining what format of
4110 dependencies will be used:
4112 1. If a static library is being produced, all upstream dependencies are
4113 required to be available in `rlib` formats. This requirement stems from the
4114 reason that a dynamic library cannot be converted into a static format.
4116 Note that it is impossible to link in native dynamic dependencies to a static
4117 library, and in this case warnings will be printed about all unlinked native
4118 dynamic dependencies.
4120 2. If an `rlib` file is being produced, then there are no restrictions on what
4121 format the upstream dependencies are available in. It is simply required that
4122 all upstream dependencies be available for reading metadata from.
4124 The reason for this is that `rlib` files do not contain any of their upstream
4125 dependencies. It wouldn't be very efficient for all `rlib` files to contain a
4126 copy of `libstd.rlib`!
4128 3. If an executable is being produced and the `-C prefer-dynamic` flag is not
4129 specified, then dependencies are first attempted to be found in the `rlib`
4130 format. If some dependencies are not available in an rlib format, then
4131 dynamic linking is attempted (see below).
4133 4. If a dynamic library or an executable that is being dynamically linked is
4134 being produced, then the compiler will attempt to reconcile the available
4135 dependencies in either the rlib or dylib format to create a final product.
4137 A major goal of the compiler is to ensure that a library never appears more
4138 than once in any artifact. For example, if dynamic libraries B and C were
4139 each statically linked to library A, then a crate could not link to B and C
4140 together because there would be two copies of A. The compiler allows mixing
4141 the rlib and dylib formats, but this restriction must be satisfied.
4143 The compiler currently implements no method of hinting what format a library
4144 should be linked with. When dynamically linking, the compiler will attempt to
4145 maximize dynamic dependencies while still allowing some dependencies to be
4146 linked in via an rlib.
4148 For most situations, having all libraries available as a dylib is recommended
4149 if dynamically linking. For other situations, the compiler will emit a
4150 warning if it is unable to determine which formats to link each library with.
4152 In general, `--crate-type=bin` or `--crate-type=lib` should be sufficient for
4153 all compilation needs, and the other options are just available if more
4154 fine-grained control is desired over the output format of a Rust crate.
4158 The runtime contains a system for directing [logging
4159 expressions](#logging-expressions) to a logging console and/or internal logging
4160 buffers. Logging can be enabled per module.
4162 Logging output is enabled by setting the `RUST_LOG` environment
4163 variable. `RUST_LOG` accepts a logging specification made up of a
4164 comma-separated list of paths, with optional log levels. For each
4165 module containing log expressions, if `RUST_LOG` contains the path to
4166 that module or a parent of that module, then logs of the appropriate
4167 level will be output to the console.
4169 The path to a module consists of the crate name, any parent modules,
4170 then the module itself, all separated by double colons (`::`). The
4171 optional log level can be appended to the module path with an equals
4172 sign (`=`) followed by the log level, from 1 to 4, inclusive. Level 1
4173 is the error level, 2 is warning, 3 info, and 4 debug. You can also
4174 use the symbolic constants `error`, `warn`, `info`, and `debug`. Any
4175 logs less than or equal to the specified level will be output. If not
4176 specified then log level 4 is assumed. Debug messages can be omitted
4177 by passing `--cfg ndebug` to `rustc`.
4179 As an example, to see all the logs generated by the compiler, you would set
4180 `RUST_LOG` to `rustc`, which is the crate name (as specified in its `crate_id`
4181 [attribute](#attributes)). To narrow down the logs to just crate resolution,
4182 you would set it to `rustc::metadata::creader`. To see just error logging
4185 Note that when compiling source files that don't specify a
4186 crate name the crate is given a default name that matches the source file,
4187 with the extension removed. In that case, to turn on logging for a program
4188 compiled from, e.g. `helloworld.rs`, `RUST_LOG` should be set to `helloworld`.
4190 #### Logging Expressions
4192 Rust provides several macros to log information. Here's a simple Rust program
4193 that demonstrates all four of them:
4197 #[phase(plugin, link)] extern crate log;
4200 error!("This is an error log")
4201 warn!("This is a warn log")
4202 info!("this is an info log")
4203 debug!("This is a debug log")
4207 These four log levels correspond to levels 1-4, as controlled by `RUST_LOG`:
4210 $ RUST_LOG=rust=3 ./rust
4211 This is an error log
4216 # Appendix: Rationales and design tradeoffs
4220 # Appendix: Influences and further references
4224 > The essential problem that must be solved in making a fault-tolerant
4225 > software system is therefore that of fault-isolation. Different programmers
4226 > will write different modules, some modules will be correct, others will have
4227 > errors. We do not want the errors in one module to adversely affect the
4228 > behaviour of a module which does not have any errors.
4230 > — Joe Armstrong
4232 > In our approach, all data is private to some process, and processes can
4233 > only communicate through communications channels. *Security*, as used
4234 > in this paper, is the property which guarantees that processes in a system
4235 > cannot affect each other except by explicit communication.
4237 > When security is absent, nothing which can be proven about a single module
4238 > in isolation can be guaranteed to hold when that module is embedded in a
4241 > — Robert Strom and Shaula Yemini
4243 > Concurrent and applicative programming complement each other. The
4244 > ability to send messages on channels provides I/O without side effects,
4245 > while the avoidance of shared data helps keep concurrent processes from
4250 Rust is not a particularly original language. It may however appear unusual
4251 by contemporary standards, as its design elements are drawn from a number of
4252 "historical" languages that have, with a few exceptions, fallen out of
4253 favour. Five prominent lineages contribute the most, though their influences
4254 have come and gone during the course of Rust's development:
4256 * The NIL (1981) and Hermes (1990) family. These languages were developed by
4257 Robert Strom, Shaula Yemini, David Bacon and others in their group at IBM
4258 Watson Research Center (Yorktown Heights, NY, USA).
4260 * The Erlang (1987) language, developed by Joe Armstrong, Robert Virding, Claes
4261 Wikström, Mike Williams and others in their group at the Ericsson Computer
4262 Science Laboratory (Älvsjö, Stockholm, Sweden) .
4264 * The Sather (1990) language, developed by Stephen Omohundro, Chu-Cheow Lim,
4265 Heinz Schmidt and others in their group at The International Computer
4266 Science Institute of the University of California, Berkeley (Berkeley, CA,
4269 * The Newsqueak (1988), Alef (1995), and Limbo (1996) family. These
4270 languages were developed by Rob Pike, Phil Winterbottom, Sean Dorward and
4271 others in their group at Bell Labs Computing Sciences Research Center
4272 (Murray Hill, NJ, USA).
4274 * The Napier (1985) and Napier88 (1988) family. These languages were
4275 developed by Malcolm Atkinson, Ron Morrison and others in their group at
4276 the University of St. Andrews (St. Andrews, Fife, UK).
4278 Additional specific influences can be seen from the following languages:
4280 * The structural algebraic types and compilation manager of SML.
4281 * The attribute and assembly systems of C#.
4282 * The references and deterministic destructor system of C++.
4283 * The memory region systems of the ML Kit and Cyclone.
4284 * The typeclass system of Haskell.
4285 * The lexical identifier rule of Python.
4286 * The block syntax of Ruby.
4288 [ffi]: guide-ffi.html