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.
384 num_lit : nonzero_dec [ dec_digit | '_' ] * num_suffix ?
385 | '0' [ [ dec_digit | '_' ] * num_suffix ?
386 | 'b' [ '1' | '0' | '_' ] + int_suffix ?
387 | 'o' [ oct_digit | '_' ] + int_suffix ?
388 | 'x' [ hex_digit | '_' ] + int_suffix ? ] ;
390 num_suffix : int_suffix | float_suffix ;
392 int_suffix : 'u' int_suffix_size ?
393 | 'i' int_suffix_size ? ;
394 int_suffix_size : [ '8' | '1' '6' | '3' '2' | '6' '4' ] ;
396 float_suffix : [ exponent | '.' dec_lit exponent ? ] ? float_suffix_ty ? ;
397 float_suffix_ty : 'f' [ '3' '2' | '6' '4' ] ;
398 exponent : ['E' | 'e'] ['-' | '+' ] ? dec_lit ;
399 dec_lit : [ dec_digit | '_' ] + ;
402 A _number literal_ is either an _integer literal_ or a _floating-point
403 literal_. The grammar for recognizing the two kinds of literals is mixed,
404 as they are differentiated by suffixes.
406 ##### Integer literals
408 An _integer literal_ has one of four forms:
410 * A _decimal literal_ starts with a *decimal digit* and continues with any
411 mixture of *decimal digits* and _underscores_.
412 * A _hex literal_ starts with the character sequence `U+0030` `U+0078`
413 (`0x`) and continues as any mixture hex digits and underscores.
414 * An _octal literal_ starts with the character sequence `U+0030` `U+006F`
415 (`0o`) and continues as any mixture octal digits and underscores.
416 * A _binary literal_ starts with the character sequence `U+0030` `U+0062`
417 (`0b`) and continues as any mixture binary digits and underscores.
419 An integer literal may be followed (immediately, without any spaces) by an
420 _integer suffix_, which changes the type of the literal. There are two kinds
421 of integer literal suffix:
423 * The `i` and `u` suffixes give the literal type `int` or `uint`,
425 * Each of the signed and unsigned machine types `u8`, `i8`,
426 `u16`, `i16`, `u32`, `i32`, `u64` and `i64`
427 give the literal the corresponding machine type.
429 The type of an _unsuffixed_ integer literal is determined by type inference.
430 If an integer type can be _uniquely_ determined from the surrounding program
431 context, the unsuffixed integer literal has that type. If the program context
432 underconstrains the type, the unsuffixed integer literal's type is `int`; if
433 the program context overconstrains the type, it is considered a static type
436 Examples of integer literals of various forms:
439 123; 0xff00; // type determined by program context
440 // defaults to int in absence of type
446 0o70_i16; // type i16
447 0b1111_1111_1001_0000_i32; // type i32
450 ##### Floating-point literals
452 A _floating-point literal_ has one of two forms:
454 * Two _decimal literals_ separated by a period
455 character `U+002E` (`.`), with an optional _exponent_ trailing after the
456 second decimal literal.
457 * A single _decimal literal_ followed by an _exponent_.
459 By default, a floating-point literal has a generic type, but will fall back to
460 `f64`. A floating-point literal may be followed (immediately, without any
461 spaces) by a _floating-point suffix_, which changes the type of the literal.
462 There are two floating-point suffixes: `f32`, and `f64` (the 32-bit and 64-bit
463 floating point types).
465 Examples of floating-point literals of various forms:
471 12E+99_f64; // type f64
474 ##### Unit and boolean literals
476 The _unit value_, the only value of the type that has the same name, is written as `()`.
477 The two values of the boolean type are written `true` and `false`.
483 | '#' | '[' | ']' | '(' | ')' | '{' | '}'
487 Symbols are a general class of printable [token](#tokens) that play structural
488 roles in a variety of grammar productions. They are catalogued here for
489 completeness as the set of remaining miscellaneous printable tokens that do not
490 otherwise appear as [unary operators](#unary-operator-expressions), [binary
491 operators](#binary-operator-expressions), or [keywords](#keywords).
497 expr_path : [ "::" ] ident [ "::" expr_path_tail ] + ;
498 expr_path_tail : '<' type_expr [ ',' type_expr ] + '>'
501 type_path : ident [ type_path_tail ] + ;
502 type_path_tail : '<' type_expr [ ',' type_expr ] + '>'
506 A _path_ is a sequence of one or more path components _logically_ separated by
507 a namespace qualifier (`::`). If a path consists of only one component, it may
508 refer to either an [item](#items) or a [slot](#memory-slots) in a local
509 control scope. If a path has multiple components, it refers to an item.
511 Every item has a _canonical path_ within its crate, but the path naming an
512 item is only meaningful within a given crate. There is no global namespace
513 across crates; an item's canonical path merely identifies it within the crate.
515 Two examples of simple paths consisting of only identifier components:
522 Path components are usually [identifiers](#identifiers), but the trailing
523 component of a path may be an angle-bracket-enclosed list of type
524 arguments. In [expression](#expressions) context, the type argument list is
525 given after a final (`::`) namespace qualifier in order to disambiguate it
526 from a relational expression involving the less-than symbol (`<`). In type
527 expression context, the final namespace qualifier is omitted.
529 Two examples of paths with type arguments:
532 # struct HashMap<K, V>;
534 # fn id<T>(t: T) -> T { t }
535 type T = HashMap<int,String>; // Type arguments used in a type expression
536 let x = id::<int>(10); // Type arguments used in a call expression
540 Paths can be denoted with various leading qualifiers to change the meaning of
543 * Paths starting with `::` are considered to be global paths where the
544 components of the path start being resolved from the crate root. Each
545 identifier in the path must resolve to an item.
553 ::a::foo(); // call a's foo function
559 * Paths starting with the keyword `super` begin resolution relative to the
560 parent module. Each further identifier must resolve to an item
568 super::a::foo(); // call a's foo function
574 * Paths starting with the keyword `self` begin resolution relative to the
575 current module. Each further identifier must resolve to an item.
587 A number of minor features of Rust are not central enough to have their own
588 syntax, and yet are not implementable as functions. Instead, they are given
589 names, and invoked through a consistent syntax: `name!(...)`. Examples
592 * `format!` : format data into a string
593 * `env!` : look up an environment variable's value at compile time
594 * `file!`: return the path to the file being compiled
595 * `stringify!` : pretty-print the Rust expression given as an argument
596 * `include!` : include the Rust expression in the given file
597 * `include_str!` : include the contents of the given file as a string
598 * `include_bin!` : include the contents of the given file as a binary blob
599 * `error!`, `warn!`, `info!`, `debug!` : provide diagnostic information.
601 All of the above extensions are expressions with values.
606 expr_macro_rules : "macro_rules" '!' ident '(' macro_rule * ')' ;
607 macro_rule : '(' matcher * ')' "=>" '(' transcriber * ')' ';' ;
608 matcher : '(' matcher * ')' | '[' matcher * ']'
609 | '{' matcher * '}' | '$' ident ':' ident
610 | '$' '(' matcher * ')' sep_token? [ '*' | '+' ]
611 | non_special_token ;
612 transcriber : '(' transcriber * ')' | '[' transcriber * ']'
613 | '{' transcriber * '}' | '$' ident
614 | '$' '(' transcriber * ')' sep_token? [ '*' | '+' ]
615 | non_special_token ;
618 User-defined syntax extensions are called "macros",
619 and the `macro_rules` syntax extension defines them.
620 Currently, user-defined macros can expand to expressions, statements, or items.
622 (A `sep_token` is any token other than `*` and `+`.
623 A `non_special_token` is any token other than a delimiter or `$`.)
625 The macro expander looks up macro invocations by name,
626 and tries each macro rule in turn.
627 It transcribes the first successful match.
628 Matching and transcription are closely related to each other,
629 and we will describe them together.
633 The macro expander matches and transcribes every token that does not begin with a `$` literally, including delimiters.
634 For parsing reasons, delimiters must be balanced, but they are otherwise not special.
636 In the matcher, `$` _name_ `:` _designator_ matches the nonterminal in the
637 Rust syntax named by _designator_. Valid designators are `item`, `block`,
638 `stmt`, `pat`, `expr`, `ty` (type), `ident`, `path`, `matchers` (lhs of the `=>` in macro rules),
639 `tt` (rhs of the `=>` in macro rules). In the transcriber, the designator is already known, and so only
640 the name of a matched nonterminal comes after the dollar sign.
642 In both the matcher and transcriber, the Kleene star-like operator indicates repetition.
643 The Kleene star operator consists of `$` and parens, optionally followed by a separator token, followed by `*` or `+`.
644 `*` means zero or more repetitions, `+` means at least one repetition.
645 The parens are not matched or transcribed.
646 On the matcher side, a name is bound to _all_ of the names it
647 matches, in a structure that mimics the structure of the repetition
648 encountered on a successful match. The job of the transcriber is to sort that
651 The rules for transcription of these repetitions are called "Macro By Example".
652 Essentially, one "layer" of repetition is discharged at a time, and all of
653 them must be discharged by the time a name is transcribed. Therefore,
654 `( $( $i:ident ),* ) => ( $i )` is an invalid macro, but
655 `( $( $i:ident ),* ) => ( $( $i:ident ),* )` is acceptable (if trivial).
657 When Macro By Example encounters a repetition, it examines all of the `$`
658 _name_ s that occur in its body. At the "current layer", they all must repeat
659 the same number of times, so
660 ` ( $( $i:ident ),* ; $( $j:ident ),* ) => ( $( ($i,$j) ),* )` is valid if
661 given the argument `(a,b,c ; d,e,f)`, but not `(a,b,c ; d,e)`. The repetition
662 walks through the choices at that layer in lockstep, so the former input
663 transcribes to `( (a,d), (b,e), (c,f) )`.
665 Nested repetitions are allowed.
667 ### Parsing limitations
669 The parser used by the macro system is reasonably powerful, but the parsing of
670 Rust syntax is restricted in two ways:
672 1. The parser will always parse as much as possible. If it attempts to match
673 `$i:expr [ , ]` against `8 [ , ]`, it will attempt to parse `i` as an array
674 index operation and fail. Adding a separator can solve this problem.
675 2. The parser must have eliminated all ambiguity by the time it reaches a `$` _name_ `:` _designator_.
676 This requirement most often affects name-designator pairs when they occur at the beginning of, or immediately after, a `$(...)*`; requiring a distinctive token in front can solve the problem.
678 ## Syntax extensions useful for the macro author
680 * `log_syntax!` : print out the arguments at compile time
681 * `trace_macros!` : supply `true` or `false` to enable or disable macro expansion logging
682 * `stringify!` : turn the identifier argument into a string literal
683 * `concat!` : concatenates a comma-separated list of literals
684 * `concat_idents!` : create a new identifier by concatenating the arguments
686 # Crates and source files
688 Rust is a *compiled* language.
689 Its semantics obey a *phase distinction* between compile-time and run-time.
690 Those semantic rules that have a *static interpretation* govern the success or failure of compilation.
691 We refer to these rules as "static semantics".
692 Semantic rules called "dynamic semantics" govern the behavior of programs at run-time.
693 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.
695 The compilation model centres on artifacts called _crates_.
696 Each compilation processes a single crate in source form, and if successful,
697 produces a single crate in binary form: either an executable or a
698 library.[^cratesourcefile]
700 [^cratesourcefile]: A crate is somewhat analogous to an *assembly* in the
701 ECMA-335 CLI model, a *library* in the SML/NJ Compilation Manager, a *unit*
702 in the Owens and Flatt module system, or a *configuration* in Mesa.
704 A _crate_ is a unit of compilation and linking, as well as versioning, distribution and runtime loading.
705 A crate contains a _tree_ of nested [module](#modules) scopes.
706 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.
708 The Rust compiler is always invoked with a single source file as input, and always produces a single output crate.
709 The processing of that source file may result in other source files being loaded as modules.
710 Source files have the extension `.rs`.
712 A Rust source file describes a module, the name and
713 location of which — in the module tree of the current crate — are defined
714 from outside the source file: either by an explicit `mod_item` in
715 a referencing source file, or by the name of the crate itself.
717 Each source file contains a sequence of zero or more `item` definitions,
718 and may optionally begin with any number of `attributes` that apply to the containing module.
719 Attributes on the anonymous crate module define important metadata that influences
720 the behavior of the compiler.
723 # #![allow(unused_attribute)]
725 #![crate_id = "projx#2.5"]
727 // Additional metadata attributes
728 #![desc = "Project X"]
730 #![comment = "This is a comment on Project X."]
732 // Specify the output type
733 #![crate_type = "lib"]
736 #![warn(non_camel_case_types)]
739 A crate that contains a `main` function can be compiled to an executable.
740 If a `main` function is present, its return type must be [`unit`](#primitive-types) and it must take no arguments.
742 # Items and attributes
744 Crates contain [items](#items),
745 each of which may have some number of [attributes](#attributes) attached to it.
750 item : mod_item | fn_item | type_item | struct_item | enum_item
751 | static_item | trait_item | impl_item | extern_block ;
754 An _item_ is a component of a crate; some module items can be defined in crate
755 files, but most are defined in source files. Items are organized within a
756 crate by a nested set of [modules](#modules). Every crate has a single
757 "outermost" anonymous module; all further items within the crate have
758 [paths](#paths) within the module tree of the crate.
760 Items are entirely determined at compile-time, generally remain fixed during
761 execution, and may reside in read-only memory.
763 There are several kinds of item:
765 * [modules](#modules)
766 * [functions](#functions)
767 * [type definitions](#type-definitions)
768 * [structures](#structures)
769 * [enumerations](#enumerations)
770 * [static items](#static-items)
772 * [implementations](#implementations)
774 Some items form an implicit scope for the declaration of sub-items. In other
775 words, within a function or module, declarations of items can (in many cases)
776 be mixed with the statements, control blocks, and similar artifacts that
777 otherwise compose the item body. The meaning of these scoped items is the same
778 as if the item was declared outside the scope — it is still a static item —
779 except that the item's *path name* within the module namespace is qualified by
780 the name of the enclosing item, or is private to the enclosing item (in the
782 The grammar specifies the exact locations in which sub-item declarations may appear.
786 All items except modules may be *parameterized* by type. Type parameters are
787 given as a comma-separated list of identifiers enclosed in angle brackets
788 (`<...>`), after the name of the item and before its definition.
789 The type parameters of an item are considered "part of the name", not part of the type of the item.
790 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.
791 In practice, the type-inference system can usually infer such argument types from context.
792 There are no general type-parametric types, only type-parametric items.
793 That is, Rust has no notion of type abstraction: there are no first-class "forall" types.
798 mod_item : "mod" ident ( ';' | '{' mod '}' );
799 mod : [ view_item | item ] * ;
802 A module is a container for zero or more [view items](#view-items) and zero or
803 more [items](#items). The view items manage the visibility of the items
804 defined within the module, as well as the visibility of names from outside the
805 module when referenced from inside the module.
807 A _module item_ is a module, surrounded in braces, named, and prefixed with
808 the keyword `mod`. A module item introduces a new, named module into the tree
809 of modules making up a crate. Modules can nest arbitrarily.
811 An example of a module:
815 type Complex = (f64, f64);
816 fn sin(f: f64) -> f64 {
820 fn cos(f: f64) -> f64 {
824 fn tan(f: f64) -> f64 {
831 Modules and types share the same namespace.
832 Declaring a named type that has the same name as a module in scope is forbidden:
833 that is, a type definition, trait, struct, enumeration, or type parameter
834 can't shadow the name of a module in scope, or vice versa.
836 A module without a body is loaded from an external file, by default with the same
837 name as the module, plus the `.rs` extension.
838 When a nested submodule is loaded from an external file,
839 it is loaded from a subdirectory path that mirrors the module hierarchy.
842 // Load the `vec` module from `vec.rs`
846 // Load the `local_data` module from `task/local_data.rs`
851 The directories and files used for loading external file modules can be influenced
852 with the `path` attribute.
855 #[path = "task_files"]
857 // Load the `local_data` module from `task_files/tls.rs`
866 view_item : extern_crate_decl | use_decl ;
869 A view item manages the namespace of a module.
870 View items do not define new items, but rather, simply change other items' visibility.
871 There are several kinds of view item:
873 * [`extern crate` declarations](#extern-crate-declarations)
874 * [`use` declarations](#use-declarations)
876 ##### Extern crate declarations
879 extern_crate_decl : "extern" "crate" ident [ '(' link_attrs ')' ] ? [ '=' string_lit ] ? ;
880 link_attrs : link_attr [ ',' link_attrs ] + ;
881 link_attr : ident '=' literal ;
884 An _`extern crate` declaration_ specifies a dependency on an external crate.
885 The external crate is then bound into the declaring scope as the `ident` provided
886 in the `extern_crate_decl`.
888 The external crate is resolved to a specific `soname` at compile time, and a
889 runtime linkage requirement to that `soname` is passed to the linker for
890 loading at runtime. The `soname` is resolved at compile time by scanning the
891 compiler's library path and matching the optional `crateid` provided as a string literal
892 against the `crateid` attributes that were declared on the external crate when
893 it was compiled. If no `crateid` is provided, a default `name` attribute is
894 assumed, equal to the `ident` given in the `extern_crate_decl`.
896 Four examples of `extern crate` declarations:
901 extern crate std; // equivalent to: extern crate std = "std";
903 extern crate ruststd = "std"; // linking to 'std' under another name
905 extern crate foo = "some/where/rust-foo#foo:1.0"; // a full crate ID for external tools
908 ##### Use declarations
911 use_decl : "pub" ? "use" [ ident '=' path
914 path_glob : ident [ "::" [ path_glob
916 | '{' ident [ ',' ident ] * '}' ;
919 A _use declaration_ creates one or more local name bindings synonymous
920 with some other [path](#paths).
921 Usually a `use` declaration is used to shorten the path required to refer to a
922 module item. These declarations may appear at the top of [modules](#modules) and
925 *Note*: Unlike in many languages,
926 `use` declarations in Rust do *not* declare linkage dependency with external crates.
927 Rather, [`extern crate` declarations](#extern-crate-declarations) declare linkage dependencies.
929 Use declarations support a number of convenient shortcuts:
931 * Rebinding the target name as a new local name, using the syntax `use x = p::q::r;`.
932 * Simultaneously binding a list of paths differing only in their final element,
933 using the glob-like brace syntax `use a::b::{c,d,e,f};`
934 * Binding all paths matching a given prefix, using the asterisk wildcard syntax `use a::b::*;`
936 An example of `use` declarations:
939 use std::iter::range_step;
940 use std::option::{Some, None};
945 // Equivalent to 'std::iter::range_step(0, 10, 2);'
946 range_step(0, 10, 2);
948 // Equivalent to 'foo(vec![std::option::Some(1.0), std::option::None]);'
949 foo(vec![Some(1.0), None]);
953 Like items, `use` declarations are private to the containing module, by default.
954 Also like items, a `use` declaration can be public, if qualified by the `pub` keyword.
955 Such a `use` declaration serves to _re-export_ a name.
956 A public `use` declaration can therefore _redirect_ some public name to a different target definition:
957 even a definition with a private canonical path, inside a different module.
958 If a sequence of such redirections form a cycle or cannot be resolved unambiguously,
959 they represent a compile-time error.
961 An example of re-exporting:
966 pub use quux::foo::{bar, baz};
975 In this example, the module `quux` re-exports two public names defined in `foo`.
977 Also note that the paths contained in `use` items are relative to the crate root.
978 So, in the previous example, the `use` refers to `quux::foo::{bar, baz}`, and not simply to `foo::{bar, baz}`.
979 This also means that top-level module declarations should be at the crate root if direct usage
980 of the declared modules within `use` items is desired. It is also possible to use `self` and `super`
981 at the beginning of a `use` item to refer to the current and direct parent modules respectively.
982 All rules regarding accessing declared modules in `use` declarations applies to both module declarations
983 and `extern crate` declarations.
985 An example of what will and will not work for `use` items:
988 # #![allow(unused_imports)]
989 use foo::native::start; // good: foo is at the root of the crate
990 use foo::baz::foobaz; // good: foo is at the root of the crate
995 use foo::native::start; // good: foo is at crate root
996 // use native::start; // bad: native is not at the crate root
997 use self::baz::foobaz; // good: self refers to module 'foo'
998 use foo::bar::foobar; // good: foo is at crate root
1005 use super::bar::foobar; // good: super refers to module 'foo'
1015 A _function item_ defines a sequence of [statements](#statements) and an optional final [expression](#expressions), along with a name and a set of parameters.
1016 Functions are declared with the keyword `fn`.
1017 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.
1019 A function may also be copied into a first class *value*, in which case the
1020 value has the corresponding [*function type*](#function-types), and can be
1021 used otherwise exactly as a function item (with a minor additional cost of
1022 calling the function indirectly).
1024 Every control path in a function logically ends with a `return` expression or a
1025 diverging expression. If the outermost block of a function has a
1026 value-producing expression in its final-expression position, that expression
1027 is interpreted as an implicit `return` expression applied to the
1030 An example of a function:
1033 fn add(x: int, y: int) -> int {
1038 As with `let` bindings, function arguments are irrefutable patterns,
1039 so any pattern that is valid in a let binding is also valid as an argument.
1042 fn first((value, _): (int, int)) -> int { value }
1046 #### Generic functions
1048 A _generic function_ allows one or more _parameterized types_ to
1049 appear in its signature. Each type parameter must be explicitly
1050 declared, in an angle-bracket-enclosed, comma-separated list following
1054 fn iter<T>(seq: &[T], f: |T|) {
1055 for elt in seq.iter() { f(elt); }
1057 fn map<T, U>(seq: &[T], f: |T| -> U) -> Vec<U> {
1058 let mut acc = vec![];
1059 for elt in seq.iter() { acc.push(f(elt)); }
1064 Inside the function signature and body, the name of the type parameter
1065 can be used as a type name.
1067 When a generic function is referenced, its type is instantiated based
1068 on the context of the reference. For example, calling the `iter`
1069 function defined above on `[1, 2]` will instantiate type parameter `T`
1070 with `int`, and require the closure parameter to have type
1073 The type parameters can also be explicitly supplied in a trailing
1074 [path](#paths) component after the function name. This might be necessary
1075 if there is not sufficient context to determine the type parameters. For
1076 example, `mem::size_of::<u32>() == 4`.
1078 Since a parameter type is opaque to the generic function, the set of
1079 operations that can be performed on it is limited. Values of parameter
1080 type can only be moved, not copied.
1083 fn id<T>(x: T) -> T { x }
1086 Similarly, [trait](#traits) bounds can be specified for type
1087 parameters to allow methods with that trait to be called on values
1093 Unsafe operations are those that potentially violate the memory-safety guarantees of Rust's static semantics.
1095 The following language level features cannot be used in the safe subset of Rust:
1097 - Dereferencing a [raw pointer](#pointer-types).
1098 - Reading or writing a [mutable static variable](#mutable-statics).
1099 - Calling an unsafe function (including an intrinsic or foreign function).
1101 ##### Unsafe functions
1103 Unsafe functions are functions that are not safe in all contexts and/or for all possible inputs.
1104 Such a function must be prefixed with the keyword `unsafe`.
1108 A block of code can also be prefixed with the `unsafe` keyword, to permit calling `unsafe` functions
1109 or dereferencing raw pointers within a safe function.
1111 When a programmer has sufficient conviction that a sequence of potentially unsafe operations is
1112 actually safe, they can encapsulate that sequence (taken as a whole) within an `unsafe` block. The
1113 compiler will consider uses of such code safe, in the surrounding context.
1115 Unsafe blocks are used to wrap foreign libraries, make direct use of hardware or implement features
1116 not directly present in the language. For example, Rust provides the language features necessary to
1117 implement memory-safe concurrency in the language but the implementation of tasks and message
1118 passing is in the standard library.
1120 Rust's type system is a conservative approximation of the dynamic safety requirements, so in some
1121 cases there is a performance cost to using safe code. For example, a doubly-linked list is not a
1122 tree structure and can only be represented with managed or reference-counted pointers in safe code.
1123 By using `unsafe` blocks to represent the reverse links as raw pointers, it can be implemented with
1124 only owned pointers.
1126 ##### Behavior considered unsafe
1128 This is a list of behavior which is forbidden in all Rust code. Type checking provides the guarantee
1129 that these issues are never caused by safe code. An `unsafe` block or function is responsible for
1130 never invoking this behaviour or exposing an API making it possible for it to occur in safe code.
1133 * Dereferencing a null/dangling raw pointer
1134 * Mutating an immutable value/reference
1135 * Reads of [undef](http://llvm.org/docs/LangRef.html#undefined-values) (uninitialized) memory
1136 * Breaking the [pointer aliasing rules](http://llvm.org/docs/LangRef.html#pointer-aliasing-rules)
1137 with raw pointers (a subset of the rules used by C)
1138 * Invoking undefined behavior via compiler intrinsics:
1139 * Indexing outside of the bounds of an object with `std::ptr::offset` (`offset` intrinsic), with
1140 the exception of one byte past the end which is permitted.
1141 * Using `std::ptr::copy_nonoverlapping_memory` (`memcpy32`/`memcpy64` instrinsics) on
1143 * Invalid values in primitive types, even in private fields/locals:
1144 * Dangling/null pointers in non-raw pointers, or slices
1145 * A value other than `false` (0) or `true` (1) in a `bool`
1146 * A discriminant in an `enum` not included in the type definition
1147 * A value in a `char` which is a surrogate or above `char::MAX`
1148 * non-UTF-8 byte sequences in a `str`
1150 ##### Behaviour not considered unsafe
1152 This is a list of behaviour not considered *unsafe* in Rust terms, but that may be undesired.
1155 * Reading data from private fields (`std::repr`, `format!("{:?}", x)`)
1156 * Leaks due to reference count cycles, even in the global heap
1157 * Exiting without calling destructors
1159 * Accessing/modifying the file system
1160 * Unsigned integer overflow (well-defined as wrapping)
1161 * Signed integer overflow (well-defined as two's complement representation wrapping)
1163 #### Diverging functions
1165 A special kind of function can be declared with a `!` character where the
1166 output slot type would normally be. For example:
1169 fn my_err(s: &str) -> ! {
1175 We call such functions "diverging" because they never return a value to the
1176 caller. Every control path in a diverging function must end with a
1177 `fail!()` or a call to another diverging function on every
1178 control path. The `!` annotation does *not* denote a type. Rather, the result
1179 type of a diverging function is a special type called $\bot$ ("bottom") that
1180 unifies with any type. Rust has no syntax for $\bot$.
1182 It might be necessary to declare a diverging function because as mentioned
1183 previously, the typechecker checks that every control path in a function ends
1184 with a [`return`](#return-expressions) or diverging expression. So, if `my_err`
1185 were declared without the `!` annotation, the following code would not
1189 # fn my_err(s: &str) -> ! { fail!() }
1191 fn f(i: int) -> int {
1196 my_err("Bad number!");
1201 This will not compile without the `!` annotation on `my_err`,
1202 since the `else` branch of the conditional in `f` does not return an `int`,
1203 as required by the signature of `f`.
1204 Adding the `!` annotation to `my_err` informs the typechecker that,
1205 should control ever enter `my_err`, no further type judgments about `f` need to hold,
1206 since control will never resume in any context that relies on those judgments.
1207 Thus the return type on `f` only needs to reflect the `if` branch of the conditional.
1210 #### Extern functions
1212 Extern functions are part of Rust's foreign function interface,
1213 providing the opposite functionality to [external blocks](#external-blocks).
1214 Whereas external blocks allow Rust code to call foreign code,
1215 extern functions with bodies defined in Rust code _can be called by foreign
1216 code_. They are defined in the same way as any other Rust function,
1217 except that they have the `extern` modifier.
1220 // Declares an extern fn, the ABI defaults to "C"
1221 extern fn new_int() -> int { 0 }
1223 // Declares an extern fn with "stdcall" ABI
1224 extern "stdcall" fn new_int_stdcall() -> int { 0 }
1227 Unlike normal functions, extern fns have an `extern "ABI" fn()`.
1228 This is the same type as the functions declared in an extern
1232 # extern fn new_int() -> int { 0 }
1233 let fptr: extern "C" fn() -> int = new_int;
1236 Extern functions may be called directly from Rust code as Rust uses large,
1237 contiguous stack segments like C.
1239 ### Type definitions
1241 A _type definition_ defines a new name for an existing [type](#types). Type
1242 definitions are declared with the keyword `type`. Every value has a single,
1243 specific type; the type-specified aspects of a value include:
1245 * Whether the value is composed of sub-values or is indivisible.
1246 * Whether the value represents textual or numerical information.
1247 * Whether the value represents integral or floating-point information.
1248 * The sequence of memory operations required to access the value.
1249 * The [kind](#type-kinds) of the type.
1251 For example, the type `(u8, u8)` defines the set of immutable values that are composite pairs,
1252 each containing two unsigned 8-bit integers accessed by pattern-matching and laid out in memory with the `x` component preceding the `y` component.
1256 A _structure_ is a nominal [structure type](#structure-types) defined with the keyword `struct`.
1258 An example of a `struct` item and its use:
1261 struct Point {x: int, y: int}
1262 let p = Point {x: 10, y: 11};
1266 A _tuple structure_ is a nominal [tuple type](#tuple-types), also defined with the keyword `struct`.
1270 struct Point(int, int);
1271 let p = Point(10, 11);
1272 let px: int = match p { Point(x, _) => x };
1275 A _unit-like struct_ is a structure without any fields, defined by leaving off the list of fields entirely.
1276 Such types will have a single value, just like the [unit value `()`](#unit-and-boolean-literals) of the unit type.
1281 let c = [Cookie, Cookie, Cookie, Cookie];
1284 By using the `struct_inherit` feature gate, structures may use single inheritance. A Structure may only
1285 inherit from a single other structure, called the _super-struct_. The inheriting structure (sub-struct)
1286 acts as if all fields in the super-struct were present in the sub-struct. Fields declared in a sub-struct
1287 must not have the same name as any field in any (transitive) super-struct. All fields (both declared
1288 and inherited) must be specified in any initializers. Inheritance between structures does not give
1289 subtyping or coercion. The super-struct and sub-struct must be defined in the same crate. The super-struct
1290 must be declared using the `virtual` keyword.
1294 virtual struct Sup { x: int }
1295 struct Sub : Sup { y: int }
1296 let s = Sub {x: 10, y: 11};
1302 An _enumeration_ is a simultaneous definition of a nominal [enumerated type](#enumerated-types) as well as a set of *constructors*,
1303 that can be used to create or pattern-match values of the corresponding enumerated type.
1305 Enumerations are declared with the keyword `enum`.
1307 An example of an `enum` item and its use:
1315 let mut a: Animal = Dog;
1319 Enumeration constructors can have either named or unnamed fields:
1322 # #![feature(struct_variant)]
1326 Cat { name: String, weight: f64 }
1329 let mut a: Animal = Dog("Cocoa".to_string(), 37.2);
1330 a = Cat { name: "Spotty".to_string(), weight: 2.7 };
1334 In this example, `Cat` is a _struct-like enum variant_,
1335 whereas `Dog` is simply called an enum variant.
1340 static_item : "static" ident ':' type '=' expr ';' ;
1343 A *static item* is a named _constant value_ stored in the global data section of a crate.
1344 Immutable static items are stored in the read-only data section.
1345 The constant value bound to a static item is, like all constant values, evaluated at compile time.
1346 Static items have the `static` lifetime, which outlives all other lifetimes in a Rust program.
1347 Static items are declared with the `static` keyword.
1348 A static item must have a _constant expression_ giving its definition.
1350 Static items must be explicitly typed.
1351 The type may be ```bool```, ```char```, a number, or a type derived from those primitive types.
1352 The derived types are references with the `static` lifetime,
1353 fixed-size arrays, tuples, and structs.
1356 static BIT1: uint = 1 << 0;
1357 static BIT2: uint = 1 << 1;
1359 static BITS: [uint, ..2] = [BIT1, BIT2];
1360 static STRING: &'static str = "bitstring";
1362 struct BitsNStrings<'a> {
1363 mybits: [uint, ..2],
1367 static bits_n_strings: BitsNStrings<'static> = BitsNStrings {
1373 #### Mutable statics
1375 If a static item is declared with the ```mut``` keyword, then it is allowed to
1376 be modified by the program. One of Rust's goals is to make concurrency bugs hard
1377 to run into, and this is obviously a very large source of race conditions or
1378 other bugs. For this reason, an ```unsafe``` block is required when either
1379 reading or writing a mutable static variable. Care should be taken to ensure
1380 that modifications to a mutable static are safe with respect to other tasks
1381 running in the same process.
1383 Mutable statics are still very useful, however. They can be used with C
1384 libraries and can also be bound from C libraries (in an ```extern``` block).
1387 # fn atomic_add(_: &mut uint, _: uint) -> uint { 2 }
1389 static mut LEVELS: uint = 0;
1391 // This violates the idea of no shared state, and this doesn't internally
1392 // protect against races, so this function is `unsafe`
1393 unsafe fn bump_levels_unsafe1() -> uint {
1399 // Assuming that we have an atomic_add function which returns the old value,
1400 // this function is "safe" but the meaning of the return value may not be what
1401 // callers expect, so it's still marked as `unsafe`
1402 unsafe fn bump_levels_unsafe2() -> uint {
1403 return atomic_add(&mut LEVELS, 1);
1409 A _trait_ describes a set of method types.
1411 Traits can include default implementations of methods,
1412 written in terms of some unknown [`self` type](#self-types);
1413 the `self` type may either be completely unspecified,
1414 or constrained by some other trait.
1416 Traits are implemented for specific types through separate [implementations](#implementations).
1419 # type Surface = int;
1420 # type BoundingBox = int;
1422 fn draw(&self, Surface);
1423 fn bounding_box(&self) -> BoundingBox;
1427 This defines a trait with two methods.
1428 All values that have [implementations](#implementations) of this trait in scope can have their `draw` and `bounding_box` methods called,
1429 using `value.bounding_box()` [syntax](#method-call-expressions).
1431 Type parameters can be specified for a trait to make it generic.
1432 These appear after the trait name, using the same syntax used in [generic functions](#generic-functions).
1436 fn len(&self) -> uint;
1437 fn elt_at(&self, n: uint) -> T;
1438 fn iter(&self, |T|);
1442 Generic functions may use traits as _bounds_ on their type parameters.
1443 This will have two effects: only types that have the trait may instantiate the parameter,
1444 and within the generic function,
1445 the methods of the trait can be called on values that have the parameter's type.
1449 # type Surface = int;
1450 # trait Shape { fn draw(&self, Surface); }
1451 fn draw_twice<T: Shape>(surface: Surface, sh: T) {
1457 Traits also define an [object type](#object-types) with the same name as the trait.
1458 Values of this type are created by [casting](#type-cast-expressions) pointer values
1459 (pointing to a type for which an implementation of the given trait is in scope)
1460 to pointers to the trait name, used as a type.
1464 # impl Shape for int { }
1466 let myshape: Box<Shape> = box mycircle as Box<Shape>;
1469 The resulting value is a box containing the value that was cast,
1470 along with information that identifies the methods of the implementation that was used.
1471 Values with a trait type can have [methods called](#method-call-expressions) on them,
1472 for any method in the trait,
1473 and can be used to instantiate type parameters that are bounded by the trait.
1475 Trait methods may be static,
1476 which means that they lack a `self` argument.
1477 This means that they can only be called with function call syntax (`f(x)`)
1478 and not method call syntax (`obj.f()`).
1479 The way to refer to the name of a static method is to qualify it with the trait name,
1480 treating the trait name like a module.
1485 fn from_int(n: int) -> Self;
1488 fn from_int(n: int) -> f64 { n as f64 }
1490 let x: f64 = Num::from_int(42);
1493 Traits may inherit from other traits. For example, in
1496 trait Shape { fn area() -> f64; }
1497 trait Circle : Shape { fn radius() -> f64; }
1500 the syntax `Circle : Shape` means that types that implement `Circle` must also have an implementation for `Shape`.
1501 Multiple supertraits are separated by `+`, `trait Circle : Shape + PartialEq { }`.
1502 In an implementation of `Circle` for a given type `T`, methods can refer to `Shape` methods,
1503 since the typechecker checks that any type with an implementation of `Circle` also has an implementation of `Shape`.
1505 In type-parameterized functions,
1506 methods of the supertrait may be called on values of subtrait-bound type parameters.
1507 Referring to the previous example of `trait Circle : Shape`:
1510 # trait Shape { fn area(&self) -> f64; }
1511 # trait Circle : Shape { fn radius(&self) -> f64; }
1512 fn radius_times_area<T: Circle>(c: T) -> f64 {
1513 // `c` is both a Circle and a Shape
1514 c.radius() * c.area()
1518 Likewise, supertrait methods may also be called on trait objects.
1521 # trait Shape { fn area(&self) -> f64; }
1522 # trait Circle : Shape { fn radius(&self) -> f64; }
1523 # impl Shape for int { fn area(&self) -> f64 { 0.0 } }
1524 # impl Circle for int { fn radius(&self) -> f64 { 0.0 } }
1526 let mycircle: Circle = ~mycircle as ~Circle;
1527 let nonsense = mycircle.radius() * mycircle.area();
1532 An _implementation_ is an item that implements a [trait](#traits) for a specific type.
1534 Implementations are defined with the keyword `impl`.
1537 # struct Point {x: f64, y: f64};
1538 # type Surface = int;
1539 # struct BoundingBox {x: f64, y: f64, width: f64, height: f64};
1540 # trait Shape { fn draw(&self, Surface); fn bounding_box(&self) -> BoundingBox; }
1541 # fn do_draw_circle(s: Surface, c: Circle) { }
1547 impl Shape for Circle {
1548 fn draw(&self, s: Surface) { do_draw_circle(s, *self); }
1549 fn bounding_box(&self) -> BoundingBox {
1550 let r = self.radius;
1551 BoundingBox{x: self.center.x - r, y: self.center.y - r,
1552 width: 2.0 * r, height: 2.0 * r}
1557 It is possible to define an implementation without referring to a trait.
1558 The methods in such an implementation can only be used
1559 as direct calls on the values of the type that the implementation targets.
1560 In such an implementation, the trait type and `for` after `impl` are omitted.
1561 Such implementations are limited to nominal types (enums, structs),
1562 and the implementation must appear in the same module or a sub-module as the `self` type.
1564 When a trait _is_ specified in an `impl`,
1565 all methods declared as part of the trait must be implemented,
1566 with matching types and type parameter counts.
1568 An implementation can take type parameters,
1569 which can be different from the type parameters taken by the trait it implements.
1570 Implementation parameters are written after the `impl` keyword.
1574 impl<T> Seq<T> for Vec<T> {
1577 impl Seq<bool> for u32 {
1578 /* Treat the integer as a sequence of bits */
1585 extern_block_item : "extern" '{' extern_block '}' ;
1586 extern_block : [ foreign_fn ] * ;
1589 External blocks form the basis for Rust's foreign function interface.
1590 Declarations in an external block describe symbols
1591 in external, non-Rust libraries.
1593 Functions within external blocks
1594 are declared in the same way as other Rust functions,
1595 with the exception that they may not have a body
1596 and are instead terminated by a semicolon.
1600 use libc::{c_char, FILE};
1603 fn fopen(filename: *c_char, mode: *c_char) -> *FILE;
1608 Functions within external blocks may be called by Rust code,
1609 just like functions defined in Rust.
1610 The Rust compiler automatically translates
1611 between the Rust ABI and the foreign ABI.
1613 A number of [attributes](#attributes) control the behavior of external
1616 By default external blocks assume that the library they are calling
1617 uses the standard C "cdecl" ABI. Other ABIs may be specified using
1618 an `abi` string, as shown here:
1621 // Interface to the Windows API
1622 extern "stdcall" { }
1625 The `link` attribute allows the name of the library to be specified. When
1626 specified the compiler will attempt to link against the native library of the
1630 #[link(name = "crypto")]
1634 The type of a function declared in an extern block is `extern "abi" fn(A1,
1635 ..., An) -> R`, where `A1...An` are the declared types of its arguments and
1636 `R` is the declared return type.
1638 ## Visibility and Privacy
1640 These two terms are often used interchangeably, and what they are attempting to
1641 convey is the answer to the question "Can this item be used at this location?"
1643 Rust's name resolution operates on a global hierarchy of namespaces. Each level
1644 in the hierarchy can be thought of as some item. The items are one of those
1645 mentioned above, but also include external crates. Declaring or defining a new
1646 module can be thought of as inserting a new tree into the hierarchy at the
1647 location of the definition.
1649 To control whether interfaces can be used across modules, Rust checks each use
1650 of an item to see whether it should be allowed or not. This is where privacy
1651 warnings are generated, or otherwise "you used a private item of another module
1652 and weren't allowed to."
1654 By default, everything in rust is *private*, with one exception. Enum variants
1655 in a `pub` enum are also public by default. You are allowed to alter this
1656 default visibility with the `priv` keyword. When an item is declared as `pub`,
1657 it can be thought of as being accessible to the outside world. For example:
1661 // Declare a private struct
1664 // Declare a public struct with a private field
1669 // Declare a public enum with two public variants
1671 PubliclyAccessibleState,
1672 PubliclyAccessibleState2,
1676 With the notion of an item being either public or private, Rust allows item
1677 accesses in two cases:
1679 1. If an item is public, then it can be used externally through any of its
1681 2. If an item is private, it may be accessed by the current module and its
1684 These two cases are surprisingly powerful for creating module hierarchies
1685 exposing public APIs while hiding internal implementation details. To help
1686 explain, here's a few use cases and what they would entail.
1688 * A library developer needs to expose functionality to crates which link against
1689 their library. As a consequence of the first case, this means that anything
1690 which is usable externally must be `pub` from the root down to the destination
1691 item. Any private item in the chain will disallow external accesses.
1693 * A crate needs a global available "helper module" to itself, but it doesn't
1694 want to expose the helper module as a public API. To accomplish this, the root
1695 of the crate's hierarchy would have a private module which then internally has
1696 a "public api". Because the entire crate is a descendant of the root, then the
1697 entire local crate can access this private module through the second case.
1699 * When writing unit tests for a module, it's often a common idiom to have an
1700 immediate child of the module to-be-tested named `mod test`. This module could
1701 access any items of the parent module through the second case, meaning that
1702 internal implementation details could also be seamlessly tested from the child
1705 In the second case, it mentions that a private item "can be accessed" by the
1706 current module and its descendants, but the exact meaning of accessing an item
1707 depends on what the item is. Accessing a module, for example, would mean looking
1708 inside of it (to import more items). On the other hand, accessing a function
1709 would mean that it is invoked. Additionally, path expressions and import
1710 statements are considered to access an item in the sense that the
1711 import/expression is only valid if the destination is in the current visibility
1714 Here's an example of a program which exemplifies the three cases outlined above.
1717 // This module is private, meaning that no external crate can access this
1718 // module. Because it is private at the root of this current crate, however, any
1719 // module in the crate may access any publicly visible item in this module.
1720 mod crate_helper_module {
1722 // This function can be used by anything in the current crate
1723 pub fn crate_helper() {}
1725 // This function *cannot* be used by anything else in the crate. It is not
1726 // publicly visible outside of the `crate_helper_module`, so only this
1727 // current module and its descendants may access it.
1728 fn implementation_detail() {}
1731 // This function is "public to the root" meaning that it's available to external
1732 // crates linking against this one.
1733 pub fn public_api() {}
1735 // Similarly to 'public_api', this module is public so external crates may look
1738 use crate_helper_module;
1740 pub fn my_method() {
1741 // Any item in the local crate may invoke the helper module's public
1742 // interface through a combination of the two rules above.
1743 crate_helper_module::crate_helper();
1746 // This function is hidden to any module which is not a descendant of
1748 fn my_implementation() {}
1754 fn test_my_implementation() {
1755 // Because this module is a descendant of `submodule`, it's allowed
1756 // to access private items inside of `submodule` without a privacy
1758 super::my_implementation();
1766 For a rust program to pass the privacy checking pass, all paths must be valid
1767 accesses given the two rules above. This includes all use statements,
1768 expressions, types, etc.
1770 ### Re-exporting and Visibility
1772 Rust allows publicly re-exporting items through a `pub use` directive. Because
1773 this is a public directive, this allows the item to be used in the current
1774 module through the rules above. It essentially allows public access into the
1775 re-exported item. For example, this program is valid:
1778 pub use api = self::implementation;
1780 mod implementation {
1787 This means that any external crate referencing `implementation::f` would receive
1788 a privacy violation, while the path `api::f` would be allowed.
1790 When re-exporting a private item, it can be thought of as allowing the "privacy
1791 chain" being short-circuited through the reexport instead of passing through the
1792 namespace hierarchy as it normally would.
1794 ### Glob imports and Visibility
1796 Currently glob imports are considered an "experimental" language feature. For
1797 sanity purpose along with helping the implementation, glob imports will only
1798 import public items from their destination, not private items.
1800 > **Note:** This is subject to change, glob exports may be removed entirely or
1801 > they could possibly import private items for a privacy error to later be
1802 > issued if the item is used.
1807 attribute : '#' '!' ? '[' meta_item ']' ;
1808 meta_item : ident [ '=' literal
1809 | '(' meta_seq ')' ] ? ;
1810 meta_seq : meta_item [ ',' meta_seq ] ? ;
1813 Static entities in Rust — crates, modules and items — may have _attributes_
1814 applied to them. Attributes in Rust are modeled on Attributes in ECMA-335,
1815 with the syntax coming from ECMA-334 (C#). An attribute is a general,
1816 free-form metadatum that is interpreted according to name, convention, and
1817 language and compiler version. Attributes may appear as any of:
1819 * A single identifier, the attribute name
1820 * An identifier followed by the equals sign '=' and a literal, providing a
1822 * An identifier followed by a parenthesized list of sub-attribute arguments
1824 Attributes with a bang ("!") after the hash ("#") apply to the item that the
1825 attribute is declared within. Attributes that do not have a bang after the
1826 hash apply to the item that follows the attribute.
1828 An example of attributes:
1831 // General metadata applied to the enclosing module or crate.
1834 // A function marked as a unit test
1840 // A conditionally-compiled module
1841 #[cfg(target_os="linux")]
1846 // A lint attribute used to suppress a warning/error
1847 #[allow(non_camel_case_types)]
1851 > **Note:** At some point in the future, the compiler will distinguish between
1852 > language-reserved and user-available attributes. Until then, there is
1853 > effectively no difference between an attribute handled by a loadable syntax
1854 > extension and the compiler.
1856 ### Crate-only attributes
1858 - `crate_id` - specify the this crate's crate ID.
1859 - `crate_type` - see [linkage](#linkage).
1860 - `feature` - see [compiler features](#compiler-features).
1861 - `no_main` - disable emitting the `main` symbol. Useful when some other
1862 object being linked to defines `main`.
1863 - `no_start` - disable linking to the `native` crate, which specifies the
1864 "start" language item.
1865 - `no_std` - disable linking to the `std` crate.
1866 - `no_builtins` - disable optimizing certain code patterns to invocations of
1867 library functions that are assumed to exist
1869 ### Module-only attributes
1871 - `macro_escape` - macros defined in this module will be visible in the
1872 module's parent, after this module has been included.
1873 - `no_implicit_prelude` - disable injecting `use std::prelude::*` in this
1875 - `path` - specifies the file to load the module from. `#[path="foo.rs"] mod
1876 bar;` is equivalent to `mod bar { /* contents of foo.rs */ }`. The path is
1877 taken relative to the directory that the current module is in.
1879 ### Function-only attributes
1881 - `plugin_registrar` - mark this function as the registration point for
1882 compiler plugins, such as loadable syntax extensions.
1883 - `main` - indicates that this function should be passed to the entry point,
1884 rather than the function in the crate root named `main`.
1885 - `start` - indicates that this function should be used as the entry point,
1886 overriding the "start" language item. See the "start" [language
1887 item](#language-items) for more details.
1889 ### Static-only attributes
1891 - `thread_local` - on a `static mut`, this signals that the value of this
1892 static may change depending on the current thread. The exact consequences of
1893 this are implementation-defined.
1897 On an `extern` block, the following attributes are interpreted:
1899 - `link_args` - specify arguments to the linker, rather than just the library
1900 name and type. This is feature gated and the exact behavior is
1901 implementation-defined (due to variety of linker invocation syntax).
1902 - `link` - indicate that a native library should be linked to for the
1903 declarations in this block to be linked correctly. See [external
1904 blocks](#external-blocks)
1906 On declarations inside an `extern` block, the following attributes are
1909 - `link_name` - the name of the symbol that this function or static should be
1911 - `linkage` - on a static, this specifies the [linkage
1912 type](http://llvm.org/docs/LangRef.html#linkage-types).
1914 ### Miscellaneous attributes
1916 - `link_section` - on statics and functions, this specifies the section of the
1917 object file that this item's contents will be placed into.
1918 - `macro_export` - export a macro for cross-crate usage.
1919 - `no_mangle` - on any item, do not apply the standard name mangling. Set the
1920 symbol for this item to its identifier.
1921 - `packed` - on structs or enums, eliminate any padding that would be used to
1923 - `repr` - on C-like enums, this sets the underlying type used for
1924 representation. Useful for FFI. Takes one argument, which is the primitive
1925 type this enum should be represented for, or `C`, which specifies that it
1926 should be the default `enum` size of the C ABI for that platform. Note that
1927 enum representation in C is undefined, and this may be incorrect when the C
1928 code is compiled with certain flags.
1929 - `simd` - on certain tuple structs, derive the arithmetic operators, which
1930 lower to the target's SIMD instructions, if any.
1931 - `static_assert` - on statics whose type is `bool`, terminates compilation
1932 with an error if it is not initialized to `true`.
1933 - `unsafe_destructor` - allow implementations of the "drop" language item
1934 where the type it is implemented for does not implement the "send" language
1936 - `unsafe_no_drop_flag` - on structs, remove the flag that prevents
1937 destructors from being run twice. Destructors might be run multiple times on
1938 the same object with this attribute.
1940 ### Conditional compilation
1942 Sometimes one wants to have different compiler outputs from the same code,
1943 depending on build target, such as targeted operating system, or to enable
1946 There are two kinds of configuration options, one that is either defined or not
1947 (`#[cfg(foo)]`), and the other that contains a string that can be checked
1948 against (`#[cfg(bar = "baz")]` (currently only compiler-defined configuration
1949 options can have the latter form).
1952 // The function is only included in the build when compiling for OSX
1953 #[cfg(target_os = "macos")]
1958 // This function is only included when either foo or bar is defined
1961 fn needs_foo_or_bar() {
1965 // This function is only included when compiling for a unixish OS with a 32-bit
1967 #[cfg(unix, target_word_size = "32")]
1968 fn on_32bit_unix() {
1973 This illustrates some conditional compilation can be achieved using the
1974 `#[cfg(...)]` attribute. Note that `#[cfg(foo, bar)]` is a condition that needs
1975 both `foo` and `bar` to be defined while `#[cfg(foo)] #[cfg(bar)]` only needs
1976 one of `foo` and `bar` to be defined (this resembles in the disjunctive normal
1977 form). Additionally, one can reverse a condition by enclosing it in a
1978 `not(...)`, like e. g. `#[cfg(not(target_os = "win32"))]`.
1980 The following configurations must be defined by the implementation:
1982 * `target_arch = "..."`. Target CPU architecture, such as `"x86"`, `"x86_64"`
1983 `"mips"`, or `"arm"`.
1984 * `target_endian = "..."`. Endianness of the target CPU, either `"little"` or
1986 * `target_family = "..."`. Operating system family of the target, e. g.
1987 `"unix"` or `"windows"`. The value of this configuration option is defined as
1988 a configuration itself, like `unix` or `windows`.
1989 * `target_os = "..."`. Operating system of the target, examples include
1990 `"win32"`, `"macos"`, `"linux"`, `"android"` or `"freebsd"`.
1991 * `target_word_size = "..."`. Target word size in bits. This is set to `"32"`
1992 for targets with 32-bit pointers, and likewise set to `"64"` for 64-bit
1994 * `unix`. See `target_family`.
1995 * `windows`. See `target_family`.
1997 ### Lint check attributes
1999 A lint check names a potentially undesirable coding pattern, such as
2000 unreachable code or omitted documentation, for the static entity to
2001 which the attribute applies.
2003 For any lint check `C`:
2005 * `warn(C)` warns about violations of `C` but continues compilation,
2006 * `deny(C)` signals an error after encountering a violation of `C`,
2007 * `allow(C)` overrides the check for `C` so that violations will go
2009 * `forbid(C)` is the same as `deny(C)`, but also forbids changing the lint
2012 The lint checks supported by the compiler can be found via `rustc -W help`,
2013 along with their default settings.
2017 // Missing documentation is ignored here
2018 #[allow(missing_doc)]
2019 pub fn undocumented_one() -> int { 1 }
2021 // Missing documentation signals a warning here
2022 #[warn(missing_doc)]
2023 pub fn undocumented_too() -> int { 2 }
2025 // Missing documentation signals an error here
2026 #[deny(missing_doc)]
2027 pub fn undocumented_end() -> int { 3 }
2031 This example shows how one can use `allow` and `warn` to toggle
2032 a particular check on and off.
2035 #[warn(missing_doc)]
2037 #[allow(missing_doc)]
2039 // Missing documentation is ignored here
2040 pub fn undocumented_one() -> int { 1 }
2042 // Missing documentation signals a warning here,
2043 // despite the allow above.
2044 #[warn(missing_doc)]
2045 pub fn undocumented_two() -> int { 2 }
2048 // Missing documentation signals a warning here
2049 pub fn undocumented_too() -> int { 3 }
2053 This example shows how one can use `forbid` to disallow uses
2054 of `allow` for that lint check.
2057 #[forbid(missing_doc)]
2059 // Attempting to toggle warning signals an error here
2060 #[allow(missing_doc)]
2062 pub fn undocumented_too() -> int { 2 }
2068 Some primitive Rust operations are defined in Rust code, rather than being
2069 implemented directly in C or assembly language. The definitions of these
2070 operations have to be easy for the compiler to find. The `lang` attribute
2071 makes it possible to declare these operations. For example, the `str` module
2072 in the Rust standard library defines the string equality function:
2076 pub fn eq_slice(a: &str, b: &str) -> bool {
2081 The name `str_eq` has a special meaning to the Rust compiler,
2082 and the presence of this definition means that it will use this definition
2083 when generating calls to the string equality function.
2085 A complete list of the built-in language items follows:
2087 #### Built-in Traits
2090 : Able to be sent across task boundaries.
2092 : Has a size known at compile time.
2094 : Types that do not move ownership when used by-value.
2096 : Able to be safely shared between tasks when aliased.
2102 These language items are traits:
2105 : Elements can be added (for example, integers and floats).
2107 : Elements can be subtracted.
2109 : Elements can be multiplied.
2111 : Elements have a division operation.
2113 : Elements have a remainder operation.
2115 : Elements can be negated arithmetically.
2117 : Elements can be negated logically.
2119 : Elements have an exclusive-or operation.
2121 : Elements have a bitwise `and` operation.
2123 : Elements have a bitwise `or` operation.
2125 : Elements have a left shift operation.
2127 : Elements have a right shift operation.
2129 : Elements can be indexed.
2131 : Elements can be compared for equality.
2133 : Elements have a partial ordering.
2135 : `*` can be applied, yielding a reference to another type
2137 : `*` can be applied, yielding a mutable reference to another type
2140 These are functions:
2143 : Compare two strings (`&str`) for equality.
2145 : Compare two owned strings (`String`) for equality.
2147 : Return a new unique string
2148 containing a copy of the contents of a unique string.
2153 : A type whose contents can be mutated through an immutable reference
2155 : The type returned by the `type_id` intrinsic.
2159 These types help drive the compiler's analysis
2162 : The type parameter should be considered covariant
2163 * `contravariant_type`
2164 : The type parameter should be considered contravariant
2166 : The type parameter should be considered invariant
2167 * `covariant_lifetime`
2168 : The lifetime parameter should be considered covariant
2169 * `contravariant_lifetime`
2170 : The lifetime parameter should be considered contravariant
2171 * `invariant_lifetime`
2172 : The lifetime parameter should be considered invariant
2174 : This type does not implement "send", even if eligible
2176 : This type does not implement "copy", even if eligible
2178 : This type does not implement "share", even if eligible
2180 : This type implements "managed"
2183 : Abort the program with an error.
2184 * `fail_bounds_check`
2185 : Abort the program with a bounds check error.
2187 : Allocate memory on the exchange heap.
2189 : Free memory that was allocated on the exchange heap.
2191 : Allocate memory on the managed heap.
2193 : Free memory that was allocated on the managed heap.
2195 > **Note:** This list is likely to become out of date. We should auto-generate it
2196 > from `librustc/middle/lang_items.rs`.
2198 ### Inline attributes
2200 The inline attribute is used to suggest to the compiler to perform an inline
2201 expansion and place a copy of the function or static in the caller rather than
2202 generating code to call the function or access the static where it is defined.
2204 The compiler automatically inlines functions based on internal heuristics.
2205 Incorrectly inlining functions can actually making the program slower, so it
2206 should be used with care.
2208 Immutable statics are always considered inlineable
2209 unless marked with `#[inline(never)]`.
2211 whether two different inlineable statics
2212 have the same memory address.
2214 the compiler is free
2215 to collapse duplicate inlineable statics together.
2217 `#[inline]` and `#[inline(always)]` always causes the function to be serialized
2218 into crate metadata to allow cross-crate inlining.
2220 There are three different types of inline attributes:
2222 * `#[inline]` hints the compiler to perform an inline expansion.
2223 * `#[inline(always)]` asks the compiler to always perform an inline expansion.
2224 * `#[inline(never)]` asks the compiler to never perform an inline expansion.
2228 The `deriving` attribute allows certain traits to be automatically
2229 implemented for data structures. For example, the following will
2230 create an `impl` for the `PartialEq` and `Clone` traits for `Foo`, the type
2231 parameter `T` will be given the `PartialEq` or `Clone` constraints for the
2235 #[deriving(PartialEq, Clone)]
2242 The generated `impl` for `PartialEq` is equivalent to
2245 # struct Foo<T> { a: int, b: T }
2246 impl<T: PartialEq> PartialEq for Foo<T> {
2247 fn eq(&self, other: &Foo<T>) -> bool {
2248 self.a == other.a && self.b == other.b
2251 fn ne(&self, other: &Foo<T>) -> bool {
2252 self.a != other.a || self.b != other.b
2257 Supported traits for `deriving` are:
2259 * Comparison traits: `PartialEq`, `TotalEq`, `PartialOrd`, `TotalOrd`.
2260 * Serialization: `Encodable`, `Decodable`. These require `serialize`.
2261 * `Clone`, to create `T` from `&T` via a copy.
2262 * `Hash`, to iterate over the bytes in a data type.
2263 * `Rand`, to create a random instance of a data type.
2264 * `Default`, to create an empty instance of a data type.
2265 * `Zero`, to create a zero instance of a numeric data type.
2266 * `FromPrimitive`, to create an instance from a numeric primitive.
2267 * `Show`, to format a value using the `{}` formatter.
2271 One can indicate the stability of an API using the following attributes:
2273 * `deprecated`: This item should no longer be used, e.g. it has been
2274 replaced. No guarantee of backwards-compatibility.
2275 * `experimental`: This item was only recently introduced or is
2276 otherwise in a state of flux. It may change significantly, or even
2277 be removed. No guarantee of backwards-compatibility.
2278 * `unstable`: This item is still under development, but requires more
2279 testing to be considered stable. No guarantee of backwards-compatibility.
2280 * `stable`: This item is considered stable, and will not change
2281 significantly. Guarantee of backwards-compatibility.
2282 * `frozen`: This item is very stable, and is unlikely to
2283 change. Guarantee of backwards-compatibility.
2284 * `locked`: This item will never change unless a serious bug is
2285 found. Guarantee of backwards-compatibility.
2287 These levels are directly inspired by
2288 [Node.js' "stability index"](http://nodejs.org/api/documentation.html).
2290 There are lints for disallowing items marked with certain levels:
2291 `deprecated`, `experimental` and `unstable`; the first two will warn
2292 by default. Items with not marked with a stability are considered to
2293 be unstable for the purposes of the lint. One can give an optional
2294 string that will be displayed when the lint flags the use of an item.
2299 #[deprecated="replaced by `best`"]
2301 // delete everything
2305 // delete fewer things
2314 bad(); // "warning: use of deprecated item: replaced by `best`"
2316 better(); // "warning: use of unmarked item"
2318 best(); // no warning
2322 > **Note:** Currently these are only checked when applied to
2323 > individual functions, structs, methods and enum variants, *not* to
2324 > entire modules, traits, impls or enums themselves.
2326 ### Compiler Features
2328 Certain aspects of Rust may be implemented in the compiler, but they're not
2329 necessarily ready for every-day use. These features are often of "prototype
2330 quality" or "almost production ready", but may not be stable enough to be
2331 considered a full-fledged language feature.
2333 For this reason, Rust recognizes a special crate-level attribute of the form:
2336 #![feature(feature1, feature2, feature3)]
2339 This directive informs the compiler that the feature list: `feature1`,
2340 `feature2`, and `feature3` should all be enabled. This is only recognized at a
2341 crate-level, not at a module-level. Without this directive, all features are
2342 considered off, and using the features will result in a compiler error.
2344 The currently implemented features of the reference compiler are:
2346 * `macro_rules` - The definition of new macros. This does not encompass
2347 macro-invocation, that is always enabled by default, this only
2348 covers the definition of new macros. There are currently
2349 various problems with invoking macros, how they interact with
2350 their environment, and possibly how they are used outside of
2351 location in which they are defined. Macro definitions are
2352 likely to change slightly in the future, so they are currently
2353 hidden behind this feature.
2355 * `globs` - Importing everything in a module through `*`. This is currently a
2356 large source of bugs in name resolution for Rust, and it's not clear
2357 whether this will continue as a feature or not. For these reasons,
2358 the glob import statement has been hidden behind this feature flag.
2360 * `struct_variant` - Structural enum variants (those with named fields). It is
2361 currently unknown whether this style of enum variant is as
2362 fully supported as the tuple-forms, and it's not certain
2363 that this style of variant should remain in the language.
2364 For now this style of variant is hidden behind a feature
2367 * `once_fns` - Onceness guarantees a closure is only executed once. Defining a
2368 closure as `once` is unlikely to be supported going forward. So
2369 they are hidden behind this feature until they are to be removed.
2371 * `managed_boxes` - Usage of `@` pointers is gated due to many
2372 planned changes to this feature. In the past, this has meant
2373 "a GC pointer", but the current implementation uses
2374 reference counting and will likely change drastically over
2375 time. Additionally, the `@` syntax will no longer be used to
2378 * `asm` - The `asm!` macro provides a means for inline assembly. This is often
2379 useful, but the exact syntax for this feature along with its semantics
2380 are likely to change, so this macro usage must be opted into.
2382 * `non_ascii_idents` - The compiler supports the use of non-ascii identifiers,
2383 but the implementation is a little rough around the
2384 edges, so this can be seen as an experimental feature for
2385 now until the specification of identifiers is fully
2388 * `thread_local` - The usage of the `#[thread_local]` attribute is experimental
2389 and should be seen as unstable. This attribute is used to
2390 declare a `static` as being unique per-thread leveraging
2391 LLVM's implementation which works in concert with the kernel
2392 loader and dynamic linker. This is not necessarily available
2393 on all platforms, and usage of it is discouraged (rust
2394 focuses more on task-local data instead of thread-local
2397 * `link_args` - This attribute is used to specify custom flags to the linker,
2398 but usage is strongly discouraged. The compiler's usage of the
2399 system linker is not guaranteed to continue in the future, and
2400 if the system linker is not used then specifying custom flags
2401 doesn't have much meaning.
2403 If a feature is promoted to a language feature, then all existing programs will
2404 start to receive compilation warnings about #[feature] directives which enabled
2405 the new feature (because the directive is no longer necessary). However, if
2406 a feature is decided to be removed from the language, errors will be issued (if
2407 there isn't a parser error first). The directive in this case is no longer
2408 necessary, and it's likely that existing code will break if the feature isn't
2411 If a unknown feature is found in a directive, it results in a compiler error. An
2412 unknown feature is one which has never been recognized by the compiler.
2414 # Statements and expressions
2416 Rust is _primarily_ an expression language. This means that most forms of
2417 value-producing or effect-causing evaluation are directed by the uniform
2418 syntax category of _expressions_. Each kind of expression can typically _nest_
2419 within each other kind of expression, and rules for evaluation of expressions
2420 involve specifying both the value produced by the expression and the order in
2421 which its sub-expressions are themselves evaluated.
2423 In contrast, statements in Rust serve _mostly_ to contain and explicitly
2424 sequence expression evaluation.
2428 A _statement_ is a component of a block, which is in turn a component of an
2429 outer [expression](#expressions) or [function](#functions).
2431 Rust has two kinds of statement:
2432 [declaration statements](#declaration-statements) and
2433 [expression statements](#expression-statements).
2435 ### Declaration statements
2437 A _declaration statement_ is one that introduces one or more *names* into the enclosing statement block.
2438 The declared names may denote new slots or new items.
2440 #### Item declarations
2442 An _item declaration statement_ has a syntactic form identical to an
2443 [item](#items) declaration within a module. Declaring an item — a function,
2444 enumeration, structure, type, static, trait, implementation or module — locally
2445 within a statement block is simply a way of restricting its scope to a narrow
2446 region containing all of its uses; it is otherwise identical in meaning to
2447 declaring the item outside the statement block.
2449 Note: there is no implicit capture of the function's dynamic environment when
2450 declaring a function-local item.
2452 #### Slot declarations
2455 let_decl : "let" pat [':' type ] ? [ init ] ? ';' ;
2456 init : [ '=' ] expr ;
2459 A _slot declaration_ introduces a new set of slots, given by a pattern.
2460 The pattern may be followed by a type annotation, and/or an initializer expression.
2461 When no type annotation is given, the compiler will infer the type,
2462 or signal an error if insufficient type information is available for definite inference.
2463 Any slots introduced by a slot declaration are visible from the point of declaration until the end of the enclosing block scope.
2465 ### Expression statements
2467 An _expression statement_ is one that evaluates an [expression](#expressions)
2468 and ignores its result.
2469 The type of an expression statement `e;` is always `()`, regardless of the type of `e`.
2470 As a rule, an expression statement's purpose is to trigger the effects of evaluating its expression.
2474 An expression may have two roles: it always produces a *value*, and it may have *effects*
2475 (otherwise known as "side effects").
2476 An expression *evaluates to* a value, and has effects during *evaluation*.
2477 Many expressions contain sub-expressions (operands).
2478 The meaning of each kind of expression dictates several things:
2479 * Whether or not to evaluate the sub-expressions when evaluating the expression
2480 * The order in which to evaluate the sub-expressions
2481 * How to combine the sub-expressions' values to obtain the value of the expression.
2483 In this way, the structure of expressions dictates the structure of execution.
2484 Blocks are just another kind of expression,
2485 so blocks, statements, expressions, and blocks again can recursively nest inside each other
2486 to an arbitrary depth.
2488 #### Lvalues, rvalues and temporaries
2490 Expressions are divided into two main categories: _lvalues_ and _rvalues_.
2491 Likewise within each expression, sub-expressions may occur in _lvalue context_ or _rvalue context_.
2492 The evaluation of an expression depends both on its own category and the context it occurs within.
2494 An lvalue is an expression that represents a memory location. These
2495 expressions are [paths](#path-expressions) (which refer to local
2496 variables, function and method arguments, or static variables),
2497 dereferences (`*expr`), [indexing expressions](#index-expressions)
2498 (`expr[expr]`), and [field references](#field-expressions) (`expr.f`).
2499 All other expressions are rvalues.
2501 The left operand of an [assignment](#assignment-expressions) or
2502 [compound-assignment](#compound-assignment-expressions) expression is an lvalue context,
2503 as is the single operand of a unary [borrow](#unary-operator-expressions).
2504 All other expression contexts are rvalue contexts.
2506 When an lvalue is evaluated in an _lvalue context_, it denotes a memory location;
2507 when evaluated in an _rvalue context_, it denotes the value held _in_ that memory location.
2509 When an rvalue is used in lvalue context, a temporary un-named lvalue is created and used instead.
2510 A temporary's lifetime equals the largest lifetime of any reference that points to it.
2512 #### Moved and copied types
2514 When a [local variable](#memory-slots) is used
2515 as an [rvalue](#lvalues-rvalues-and-temporaries)
2516 the variable will either be moved or copied, depending on its type.
2517 For types that contain [owning pointers](#pointer-types)
2518 or values that implement the special trait `Drop`,
2519 the variable is moved.
2520 All other types are copied.
2522 ### Literal expressions
2524 A _literal expression_ consists of one of the [literal](#literals)
2525 forms described earlier. It directly describes a number, character,
2526 string, boolean value, or the unit value.
2530 "hello"; // string type
2531 '5'; // character type
2535 ### Path expressions
2537 A [path](#paths) used as an expression context denotes either a local variable or an item.
2538 Path expressions are [lvalues](#lvalues-rvalues-and-temporaries).
2540 ### Tuple expressions
2542 Tuples are written by enclosing one or more comma-separated
2543 expressions in parentheses. They are used to create [tuple-typed](#tuple-types)
2552 ### Structure expressions
2555 struct_expr : expr_path '{' ident ':' expr
2556 [ ',' ident ':' expr ] *
2559 [ ',' expr ] * ')' |
2563 There are several forms of structure expressions.
2564 A _structure expression_ consists of the [path](#paths) of a [structure item](#structures),
2565 followed by a brace-enclosed list of one or more comma-separated name-value pairs,
2566 providing the field values of a new instance of the structure.
2567 A field name can be any identifier, and is separated from its value expression by a colon.
2568 The location denoted by a structure field is mutable if and only if the enclosing structure is mutable.
2570 A _tuple structure expression_ consists of the [path](#paths) of a [structure item](#structures),
2571 followed by a parenthesized list of one or more comma-separated expressions
2572 (in other words, the path of a structure item followed by a tuple expression).
2573 The structure item must be a tuple structure item.
2575 A _unit-like structure expression_ consists only of the [path](#paths) of a [structure item](#structures).
2577 The following are examples of structure expressions:
2580 # struct Point { x: f64, y: f64 }
2581 # struct TuplePoint(f64, f64);
2582 # mod game { pub struct User<'a> { pub name: &'a str, pub age: uint, pub score: uint } }
2583 # struct Cookie; fn some_fn<T>(t: T) {}
2584 Point {x: 10.0, y: 20.0};
2585 TuplePoint(10.0, 20.0);
2586 let u = game::User {name: "Joe", age: 35, score: 100_000};
2587 some_fn::<Cookie>(Cookie);
2590 A structure expression forms a new value of the named structure type.
2591 Note that for a given *unit-like* structure type, this will always be the same value.
2593 A structure expression can terminate with the syntax `..` followed by an expression to denote a functional update.
2594 The expression following `..` (the base) must have the same structure type as the new structure type being formed.
2595 The entire expression denotes the result of constructing a new structure
2596 (with the same type as the base expression)
2597 with the given values for the fields that were explicitly specified
2598 and the values in the base expression for all other fields.
2601 # struct Point3d { x: int, y: int, z: int }
2602 let base = Point3d {x: 1, y: 2, z: 3};
2603 Point3d {y: 0, z: 10, .. base};
2606 ### Block expressions
2609 block_expr : '{' [ view_item ] *
2610 [ stmt ';' | item ] *
2614 A _block expression_ is similar to a module in terms of the declarations that
2615 are possible. Each block conceptually introduces a new namespace scope. View
2616 items can bring new names into scopes and declared items are in scope for only
2619 A block will execute each statement sequentially, and then execute the
2620 expression (if given). If the final expression is omitted, the type and return
2621 value of the block are `()`, but if it is provided, the type and return value
2622 of the block are that of the expression itself.
2624 ### Method-call expressions
2627 method_call_expr : expr '.' ident paren_expr_list ;
2630 A _method call_ consists of an expression followed by a single dot, an identifier, and a parenthesized expression-list.
2631 Method calls are resolved to methods on specific traits,
2632 either statically dispatching to a method if the exact `self`-type of the left-hand-side is known,
2633 or dynamically dispatching if the left-hand-side expression is an indirect [object type](#object-types).
2635 ### Field expressions
2638 field_expr : expr '.' ident ;
2641 A _field expression_ consists of an expression followed by a single dot and an identifier,
2642 when not immediately followed by a parenthesized expression-list (the latter is a [method call expression](#method-call-expressions)).
2643 A field expression denotes a field of a [structure](#structure-types).
2645 ~~~~ {.ignore .field}
2648 (Struct {a: 10, b: 20}).a;
2651 A field access is an [lvalue](#lvalues-rvalues-and-temporaries) referring to the value of that field.
2652 When the type providing the field inherits mutabilty, it can be [assigned](#assignment-expressions) to.
2654 Also, if the type of the expression to the left of the dot is a pointer,
2655 it is automatically dereferenced to make the field access possible.
2657 ### Vector expressions
2660 vec_expr : '[' "mut" ? vec_elems? ']' ;
2662 vec_elems : [expr [',' expr]*] | [expr ',' ".." expr] ;
2665 A [_vector_](#vector-types) _expression_ is written by enclosing zero or
2666 more comma-separated expressions of uniform type in square brackets.
2668 In the `[expr ',' ".." expr]` form, the expression after the `".."`
2669 must be a constant expression that can be evaluated at compile time, such
2670 as a [literal](#literals) or a [static item](#static-items).
2674 ["a", "b", "c", "d"];
2675 [0, ..128]; // vector with 128 zeros
2676 [0u8, 0u8, 0u8, 0u8];
2679 ### Index expressions
2682 idx_expr : expr '[' expr ']' ;
2685 [Vector](#vector-types)-typed expressions can be indexed by writing a
2686 square-bracket-enclosed expression (the index) after them. When the
2687 vector is mutable, the resulting [lvalue](#lvalues-rvalues-and-temporaries) can be assigned to.
2689 Indices are zero-based, and may be of any integral type. Vector access
2690 is bounds-checked at run-time. When the check fails, it will put the
2691 task in a _failing state_.
2695 # task::spawn(proc() {
2698 (["a", "b"])[10]; // fails
2703 ### Unary operator expressions
2705 Rust defines six symbolic unary operators.
2706 They are all written as prefix operators,
2707 before the expression they apply to.
2710 : Negation. May only be applied to numeric types.
2712 : Dereference. When applied to a [pointer](#pointer-types) it denotes the pointed-to location.
2713 For pointers to mutable locations, the resulting [lvalue](#lvalues-rvalues-and-temporaries) can be assigned to.
2714 On non-pointer types, it calls the `deref` method of the `std::ops::Deref` trait, or the
2715 `deref_mut` method of the `std::ops::DerefMut` trait (if implemented by the type and required
2716 for an outer expression that will or could mutate the dereference), and produces the
2717 result of dereferencing the `&` or `&mut` borrowed pointer returned from the overload method.
2720 : Logical negation. On the boolean type, this flips between `true` and
2721 `false`. On integer types, this inverts the individual bits in the
2722 two's complement representation of the value.
2724 : [Boxing](#pointer-types) operators. Allocate a box to hold the value they are applied to,
2725 and store the value in it. `box` creates an owned box.
2727 : Borrow operator. Returns a reference, pointing to its operand.
2728 The operand of a borrow is statically proven to outlive the resulting pointer.
2729 If the borrow-checker cannot prove this, it is a compilation error.
2731 ### Binary operator expressions
2734 binop_expr : expr binop expr ;
2737 Binary operators expressions are given in terms of
2738 [operator precedence](#operator-precedence).
2740 #### Arithmetic operators
2742 Binary arithmetic expressions are syntactic sugar for calls to built-in traits,
2743 defined in the `std::ops` module of the `std` library.
2744 This means that arithmetic operators can be overridden for user-defined types.
2745 The default meaning of the operators on standard types is given here.
2748 : Addition and vector/string concatenation.
2749 Calls the `add` method on the `std::ops::Add` trait.
2752 Calls the `sub` method on the `std::ops::Sub` trait.
2755 Calls the `mul` method on the `std::ops::Mul` trait.
2758 Calls the `div` method on the `std::ops::Div` trait.
2761 Calls the `rem` method on the `std::ops::Rem` trait.
2763 #### Bitwise operators
2765 Like the [arithmetic operators](#arithmetic-operators), bitwise operators
2766 are syntactic sugar for calls to methods of built-in traits.
2767 This means that bitwise operators can be overridden for user-defined types.
2768 The default meaning of the operators on standard types is given here.
2772 Calls the `bitand` method of the `std::ops::BitAnd` trait.
2775 Calls the `bitor` method of the `std::ops::BitOr` trait.
2778 Calls the `bitxor` method of the `std::ops::BitXor` trait.
2780 : Logical left shift.
2781 Calls the `shl` method of the `std::ops::Shl` trait.
2783 : Logical right shift.
2784 Calls the `shr` method of the `std::ops::Shr` trait.
2786 #### Lazy boolean operators
2788 The operators `||` and `&&` may be applied to operands of boolean type.
2789 The `||` operator denotes logical 'or', and the `&&` operator denotes logical 'and'.
2790 They differ from `|` and `&` in that the right-hand operand is only evaluated
2791 when the left-hand operand does not already determine the result of the expression.
2792 That is, `||` only evaluates its right-hand operand
2793 when the left-hand operand evaluates to `false`, and `&&` only when it evaluates to `true`.
2795 #### Comparison operators
2797 Comparison operators are, like the [arithmetic operators](#arithmetic-operators),
2798 and [bitwise operators](#bitwise-operators),
2799 syntactic sugar for calls to built-in traits.
2800 This means that comparison operators can be overridden for user-defined types.
2801 The default meaning of the operators on standard types is given here.
2805 Calls the `eq` method on the `std::cmp::PartialEq` trait.
2808 Calls the `ne` method on the `std::cmp::PartialEq` trait.
2811 Calls the `lt` method on the `std::cmp::PartialOrd` trait.
2814 Calls the `gt` method on the `std::cmp::PartialOrd` trait.
2816 : Less than or equal.
2817 Calls the `le` method on the `std::cmp::PartialOrd` trait.
2819 : Greater than or equal.
2820 Calls the `ge` method on the `std::cmp::PartialOrd` trait.
2822 #### Type cast expressions
2824 A type cast expression is denoted with the binary operator `as`.
2826 Executing an `as` expression casts the value on the left-hand side to the type
2827 on the right-hand side.
2829 A numeric value can be cast to any numeric type.
2830 A raw pointer value can be cast to or from any integral type or raw pointer type.
2831 Any other cast is unsupported and will fail to compile.
2833 An example of an `as` expression:
2836 # fn sum(v: &[f64]) -> f64 { 0.0 }
2837 # fn len(v: &[f64]) -> int { 0 }
2839 fn avg(v: &[f64]) -> f64 {
2840 let sum: f64 = sum(v);
2841 let sz: f64 = len(v) as f64;
2846 #### Assignment expressions
2848 An _assignment expression_ consists of an [lvalue](#lvalues-rvalues-and-temporaries) expression followed by an
2849 equals sign (`=`) and an [rvalue](#lvalues-rvalues-and-temporaries) expression.
2851 Evaluating an assignment expression [either copies or moves](#moved-and-copied-types) its right-hand operand to its left-hand operand.
2860 #### Compound assignment expressions
2862 The `+`, `-`, `*`, `/`, `%`, `&`, `|`, `^`, `<<`, and `>>`
2863 operators may be composed with the `=` operator. The expression `lval
2864 OP= val` is equivalent to `lval = lval OP val`. For example, `x = x +
2865 1` may be written as `x += 1`.
2867 Any such expression always has the [`unit`](#primitive-types) type.
2869 #### Operator precedence
2871 The precedence of Rust binary operators is ordered as follows, going
2872 from strong to weak:
2874 ~~~~ {.text .precedence}
2889 Operators at the same precedence level are evaluated left-to-right. [Unary operators](#unary-operator-expressions)
2890 have the same precedence level and it is stronger than any of the binary operators'.
2892 ### Grouped expressions
2894 An expression enclosed in parentheses evaluates to the result of the enclosed
2895 expression. Parentheses can be used to explicitly specify evaluation order
2896 within an expression.
2899 paren_expr : '(' expr ')' ;
2902 An example of a parenthesized expression:
2905 let x = (2 + 3) * 4;
2909 ### Call expressions
2912 expr_list : [ expr [ ',' expr ]* ] ? ;
2913 paren_expr_list : '(' expr_list ')' ;
2914 call_expr : expr paren_expr_list ;
2917 A _call expression_ invokes a function, providing zero or more input slots and
2918 an optional reference slot to serve as the function's output, bound to the
2919 `lval` on the right hand side of the call. If the function eventually returns,
2920 then the expression completes.
2922 Some examples of call expressions:
2925 # use std::from_str::FromStr;
2926 # fn add(x: int, y: int) -> int { 0 }
2928 let x: int = add(1, 2);
2929 let pi: Option<f32> = FromStr::from_str("3.14");
2932 ### Lambda expressions
2935 ident_list : [ ident [ ',' ident ]* ] ? ;
2936 lambda_expr : '|' ident_list '|' expr ;
2939 A _lambda expression_ (sometimes called an "anonymous function expression") defines a function and denotes it as a value,
2940 in a single expression.
2941 A lambda expression is a pipe-symbol-delimited (`|`) list of identifiers followed by an expression.
2943 A lambda expression denotes a function that maps a list of parameters (`ident_list`)
2944 onto the expression that follows the `ident_list`.
2945 The identifiers in the `ident_list` are the parameters to the function.
2946 These parameters' types need not be specified, as the compiler infers them from context.
2948 Lambda expressions are most useful when passing functions as arguments to other functions,
2949 as an abbreviation for defining and capturing a separate function.
2951 Significantly, lambda expressions _capture their environment_,
2952 which regular [function definitions](#functions) do not.
2953 The exact type of capture depends on the [function type](#function-types) inferred for the lambda expression.
2954 In the simplest and least-expensive form (analogous to a ```|| { }``` expression),
2955 the lambda expression captures its environment by reference,
2956 effectively borrowing pointers to all outer variables mentioned inside the function.
2957 Alternately, the compiler may infer that a lambda expression should copy or move values (depending on their type.)
2958 from the environment into the lambda expression's captured environment.
2960 In this example, we define a function `ten_times` that takes a higher-order function argument,
2961 and call it with a lambda expression as an argument.
2964 fn ten_times(f: |int|) {
2972 ten_times(|j| println!("hello, {}", j));
2978 while_expr : "while" expr '{' block '}' ;
2981 A `while` loop begins by evaluating the boolean loop conditional expression.
2982 If the loop conditional expression evaluates to `true`, the loop body block
2983 executes and control returns to the loop conditional expression. If the loop
2984 conditional expression evaluates to `false`, the `while` expression completes.
2999 A `loop` expression denotes an infinite loop.
3002 loop_expr : [ lifetime ':' ] "loop" '{' block '}';
3005 A `loop` expression may optionally have a _label_.
3006 If a label is present,
3007 then labeled `break` and `continue` expressions nested within this loop may exit out of this loop or return control to its head.
3008 See [Break expressions](#break-expressions) and [Continue expressions](#continue-expressions).
3010 ### Break expressions
3013 break_expr : "break" [ lifetime ];
3016 A `break` expression has an optional _label_.
3017 If the label is absent, then executing a `break` expression immediately terminates the innermost loop enclosing it.
3018 It is only permitted in the body of a loop.
3019 If the label is present, then `break foo` terminates the loop with label `foo`,
3020 which need not be the innermost label enclosing the `break` expression,
3021 but must enclose it.
3023 ### Continue expressions
3026 continue_expr : "continue" [ lifetime ];
3029 A `continue` expression has an optional _label_.
3030 If the label is absent,
3031 then executing a `continue` expression immediately terminates the current iteration of the innermost loop enclosing it,
3032 returning control to the loop *head*.
3033 In the case of a `while` loop,
3034 the head is the conditional expression controlling the loop.
3035 In the case of a `for` loop, the head is the call-expression controlling the loop.
3036 If the label is present, then `continue foo` returns control to the head of the loop with label `foo`,
3037 which need not be the innermost label enclosing the `break` expression,
3038 but must enclose it.
3040 A `continue` expression is only permitted in the body of a loop.
3045 for_expr : "for" pat "in" expr '{' block '}' ;
3048 A `for` expression is a syntactic construct for looping over elements
3049 provided by an implementation of `std::iter::Iterator`.
3051 An example of a for loop over the contents of a vector:
3055 # fn bar(f: Foo) { }
3060 let v: &[Foo] = &[a, b, c];
3067 An example of a for loop over a series of integers:
3070 # fn bar(b:uint) { }
3071 for i in range(0u, 256) {
3079 if_expr : "if" expr '{' block '}'
3082 else_tail : "else" [ if_expr
3086 An `if` expression is a conditional branch in program control. The form of
3087 an `if` expression is a condition expression, followed by a consequent
3088 block, any number of `else if` conditions and blocks, and an optional
3089 trailing `else` block. The condition expressions must have type
3090 `bool`. If a condition expression evaluates to `true`, the
3091 consequent block is executed and any subsequent `else if` or `else`
3092 block is skipped. If a condition expression evaluates to `false`, the
3093 consequent block is skipped and any subsequent `else if` condition is
3094 evaluated. If all `if` and `else if` conditions evaluate to `false`
3095 then any `else` block is executed.
3097 ### Match expressions
3100 match_expr : "match" expr '{' match_arm * '}' ;
3102 match_arm : attribute * match_pat "=>" [ expr "," | '{' block '}' ] ;
3104 match_pat : pat [ '|' pat ] * [ "if" expr ] ? ;
3107 A `match` expression branches on a *pattern*. The exact form of matching that
3108 occurs depends on the pattern. Patterns consist of some combination of
3109 literals, destructured vectors or enum constructors, structures and
3110 tuples, variable binding specifications, wildcards (`..`), and placeholders
3111 (`_`). A `match` expression has a *head expression*, which is the value to
3112 compare to the patterns. The type of the patterns must equal the type of the
3115 In a pattern whose head expression has an `enum` type, a placeholder (`_`)
3116 stands for a *single* data field, whereas a wildcard `..` stands for *all* the
3117 fields of a particular variant. For example:
3120 enum List<X> { Nil, Cons(X, Box<List<X>>) }
3122 let x: List<int> = Cons(10, box Cons(11, box Nil));
3125 Cons(_, box Nil) => fail!("singleton list"),
3127 Nil => fail!("empty list")
3131 The first pattern matches lists constructed by applying `Cons` to any head
3132 value, and a tail value of `box Nil`. The second pattern matches _any_ list
3133 constructed with `Cons`, ignoring the values of its arguments. The difference
3134 between `_` and `..` is that the pattern `C(_)` is only type-correct if `C` has
3135 exactly one argument, while the pattern `C(..)` is type-correct for any enum
3136 variant `C`, regardless of how many arguments `C` has.
3138 Used inside a vector pattern, `..` stands for any number of elements. This
3139 wildcard can be used at most once for a given vector, which implies that it
3140 cannot be used to specifically match elements that are at an unknown distance
3141 from both ends of a vector, like `[.., 42, ..]`. If followed by a variable name,
3142 it will bind the corresponding slice to the variable. Example:
3145 fn is_symmetric(list: &[uint]) -> bool {
3148 [x, ..inside, y] if x == y => is_symmetric(inside),
3154 let sym = &[0, 1, 4, 2, 4, 1, 0];
3155 let not_sym = &[0, 1, 7, 2, 4, 1, 0];
3156 assert!(is_symmetric(sym));
3157 assert!(!is_symmetric(not_sym));
3161 A `match` behaves differently depending on whether or not the head expression
3162 is an [lvalue or an rvalue](#lvalues-rvalues-and-temporaries).
3163 If the head expression is an rvalue, it is
3164 first evaluated into a temporary location, and the resulting value
3165 is sequentially compared to the patterns in the arms until a match
3166 is found. The first arm with a matching pattern is chosen as the branch target
3167 of the `match`, any variables bound by the pattern are assigned to local
3168 variables in the arm's block, and control enters the block.
3170 When the head expression is an lvalue, the match does not allocate a
3171 temporary location (however, a by-value binding may copy or move from
3172 the lvalue). When possible, it is preferable to match on lvalues, as the
3173 lifetime of these matches inherits the lifetime of the lvalue, rather
3174 than being restricted to the inside of the match.
3176 An example of a `match` expression:
3179 # fn process_pair(a: int, b: int) { }
3180 # fn process_ten() { }
3182 enum List<X> { Nil, Cons(X, Box<List<X>>) }
3184 let x: List<int> = Cons(10, box Cons(11, box Nil));
3187 Cons(a, box Cons(b, _)) => {
3202 Patterns that bind variables
3203 default to binding to a copy or move of the matched value
3204 (depending on the matched value's type).
3205 This can be changed to bind to a reference by
3206 using the `ref` keyword,
3207 or to a mutable reference using `ref mut`.
3209 Subpatterns can also be bound to variables by the use of the syntax
3210 `variable @ subpattern`.
3214 enum List { Nil, Cons(uint, Box<List>) }
3216 fn is_sorted(list: &List) -> bool {
3218 Nil | Cons(_, box Nil) => true,
3219 Cons(x, ref r @ box Cons(y, _)) => (x <= y) && is_sorted(*r)
3224 let a = Cons(6, box Cons(7, box Cons(42, box Nil)));
3225 assert!(is_sorted(&a));
3230 Patterns can also dereference pointers by using the `&`,
3231 `box` or `@` symbols, as appropriate. For example, these two matches
3232 on `x: &int` are equivalent:
3236 let y = match *x { 0 => "zero", _ => "some" };
3237 let z = match x { &0 => "zero", _ => "some" };
3242 A pattern that's just an identifier, like `Nil` in the previous example,
3243 could either refer to an enum variant that's in scope, or bind a new variable.
3244 The compiler resolves this ambiguity by forbidding variable bindings that occur
3245 in `match` patterns from shadowing names of variants that are in scope.
3246 For example, wherever `List` is in scope,
3247 a `match` pattern would not be able to bind `Nil` as a new name.
3248 The compiler interprets a variable pattern `x` as a binding _only_ if there is
3249 no variant named `x` in scope.
3250 A convention you can use to avoid conflicts is simply to name variants with
3251 upper-case letters, and local variables with lower-case letters.
3253 Multiple match patterns may be joined with the `|` operator.
3254 A range of values may be specified with `..`.
3260 let message = match x {
3261 0 | 1 => "not many",
3267 Range patterns only work on scalar types
3268 (like integers and characters; not like vectors and structs, which have sub-components).
3269 A range pattern may not be a sub-range of another range pattern inside the same `match`.
3271 Finally, match patterns can accept *pattern guards* to further refine the
3272 criteria for matching a case. Pattern guards appear after the pattern and
3273 consist of a bool-typed expression following the `if` keyword. A pattern
3274 guard may refer to the variables bound within the pattern they follow.
3277 # let maybe_digit = Some(0);
3278 # fn process_digit(i: int) { }
3279 # fn process_other(i: int) { }
3281 let message = match maybe_digit {
3282 Some(x) if x < 10 => process_digit(x),
3283 Some(x) => process_other(x),
3288 ### Return expressions
3291 return_expr : "return" expr ? ;
3294 Return expressions are denoted with the keyword `return`. Evaluating a `return`
3295 expression moves its argument into the output slot of the current
3296 function, destroys the current function activation frame, and transfers
3297 control to the caller frame.
3299 An example of a `return` expression:
3302 fn max(a: int, b: int) -> int {
3314 Every slot, item and value in a Rust program has a type. The _type_ of a *value*
3315 defines the interpretation of the memory holding it.
3317 Built-in types and type-constructors are tightly integrated into the language,
3318 in nontrivial ways that are not possible to emulate in user-defined
3319 types. User-defined types have limited capabilities.
3323 The primitive types are the following:
3325 * The "unit" type `()`, having the single "unit" value `()` (occasionally called
3327 * The boolean type `bool` with values `true` and `false`.
3328 * The machine types.
3329 * The machine-dependent integer and floating-point types.
3331 [^unittype]: The "unit" value `()` is *not* a sentinel "null pointer" value for
3332 reference slots; the "unit" type is the implicit return type from functions
3333 otherwise lacking a return type, and can be used in other contexts (such as
3334 message-sending or type-parametric code) as a zero-size type.]
3338 The machine types are the following:
3340 * The unsigned word types `u8`, `u16`, `u32` and `u64`, with values drawn from
3341 the integer intervals [0, 2^8 - 1], [0, 2^16 - 1], [0, 2^32 - 1] and
3342 [0, 2^64 - 1] respectively.
3344 * The signed two's complement word types `i8`, `i16`, `i32` and `i64`, with
3345 values drawn from the integer intervals [-(2^(7)), 2^7 - 1],
3346 [-(2^(15)), 2^15 - 1], [-(2^(31)), 2^31 - 1], [-(2^(63)), 2^63 - 1]
3349 * The IEEE 754-2008 `binary32` and `binary64` floating-point types: `f32` and
3350 `f64`, respectively.
3352 #### Machine-dependent integer types
3354 The Rust type `uint` [^rustuint] is an
3355 unsigned integer type with target-machine-dependent size. Its size, in
3356 bits, is equal to the number of bits required to hold any memory address on
3359 The Rust type `int` [^rustint] is a
3360 two's complement signed integer type with target-machine-dependent size. Its
3361 size, in bits, is equal to the size of the rust type `uint` on the same target
3364 [^rustuint]: A Rust `uint` is analogous to a C99 `uintptr_t`.
3365 [^rustint]: A Rust `int` is analogous to a C99 `intptr_t`.
3369 The types `char` and `str` hold textual data.
3371 A value of type `char` is a [Unicode scalar value](
3372 http://www.unicode.org/glossary/#unicode_scalar_value)
3373 (ie. a code point that is not a surrogate),
3374 represented as a 32-bit unsigned word in the 0x0000 to 0xD7FF
3375 or 0xE000 to 0x10FFFF range.
3376 A `[char]` vector is effectively an UCS-4 / UTF-32 string.
3378 A value of type `str` is a Unicode string,
3379 represented as a vector of 8-bit unsigned bytes holding a sequence of UTF-8 codepoints.
3380 Since `str` is of unknown size, it is not a _first class_ type,
3381 but can only be instantiated through a pointer type,
3382 such as `&str` or `String`.
3386 A tuple *type* is a heterogeneous product of other types, called the *elements*
3387 of the tuple. It has no nominal name and is instead structurally typed.
3389 Tuple types and values are denoted by listing the types or values of their
3390 elements, respectively, in a parenthesized, comma-separated
3393 Because tuple elements don't have a name, they can only be accessed by pattern-matching.
3395 The members of a tuple are laid out in memory contiguously, in
3396 order specified by the tuple type.
3398 An example of a tuple type and its use:
3401 type Pair<'a> = (int, &'a str);
3402 let p: Pair<'static> = (10, "hello");
3404 assert!(b != "world");
3409 The vector type constructor represents a homogeneous array of values of a given type.
3410 A vector has a fixed size.
3411 (Operations like `vec.push` operate solely on owned vectors.)
3412 A vector type can be annotated with a _definite_ size, such as `[int, ..10]`.
3413 Such a definite-sized vector type is a first-class type, since its size is known statically.
3414 A vector without such a size is said to be of _indefinite_ size,
3415 and is therefore not a _first-class_ type.
3416 An indefinite-size vector can only be instantiated through a pointer type,
3417 such as `&[T]` or `Vec<T>`.
3418 The kind of a vector type depends on the kind of its element type,
3419 as with other simple structural types.
3421 Expressions producing vectors of definite size cannot be evaluated in a
3422 context expecting a vector of indefinite size; one must copy the
3423 definite-sized vector contents into a distinct vector of indefinite size.
3425 An example of a vector type and its use:
3428 let v: &[int] = &[7, 5, 3];
3433 All in-bounds elements of a vector are always initialized,
3434 and access to a vector is always bounds-checked.
3438 A `struct` *type* is a heterogeneous product of other types, called the *fields*
3439 of the type.[^structtype]
3441 [^structtype]: `struct` types are analogous `struct` types in C,
3442 the *record* types of the ML family,
3443 or the *structure* types of the Lisp family.
3445 New instances of a `struct` can be constructed with a [struct expression](#structure-expressions).
3447 The memory layout of a `struct` is undefined by default to allow for compiler optimziations like
3448 field reordering, but it can be fixed with the `#[repr(...)]` attribute.
3449 In either case, fields may be given in any order in a corresponding struct *expression*;
3450 the resulting `struct` value will always have the same memory layout.
3452 The fields of a `struct` may be qualified by [visibility modifiers](#re-exporting-and-visibility),
3453 to allow access to data in a structure outside a module.
3455 A _tuple struct_ type is just like a structure type, except that the fields are anonymous.
3457 A _unit-like struct_ type is like a structure type, except that it has no fields.
3458 The one value constructed by the associated [structure expression](#structure-expressions)
3459 is the only value that inhabits such a type.
3461 ### Enumerated types
3463 An *enumerated type* is a nominal, heterogeneous disjoint union type,
3464 denoted by the name of an [`enum` item](#enumerations). [^enumtype]
3466 [^enumtype]: The `enum` type is analogous to a `data` constructor declaration in
3467 ML, or a *pick ADT* in Limbo.
3469 An [`enum` item](#enumerations) declares both the type and a number of *variant constructors*,
3470 each of which is independently named and takes an optional tuple of arguments.
3472 New instances of an `enum` can be constructed by calling one of the variant constructors,
3473 in a [call expression](#call-expressions).
3475 Any `enum` value consumes as much memory as the largest variant constructor for its corresponding `enum` type.
3477 Enum types cannot be denoted *structurally* as types,
3478 but must be denoted by named reference to an [`enum` item](#enumerations).
3482 Nominal types — [enumerations](#enumerated-types) and [structures](#structure-types) — may be recursive.
3483 That is, each `enum` constructor or `struct` field may refer, directly or indirectly, to the enclosing `enum` or `struct` type itself.
3484 Such recursion has restrictions:
3486 * Recursive types must include a nominal type in the recursion
3487 (not mere [type definitions](#type-definitions),
3488 or other structural types such as [vectors](#vector-types) or [tuples](#tuple-types)).
3489 * A recursive `enum` item must have at least one non-recursive constructor
3490 (in order to give the recursion a basis case).
3491 * The size of a recursive type must be finite;
3492 in other words the recursive fields of the type must be [pointer types](#pointer-types).
3493 * Recursive type definitions can cross module boundaries, but not module *visibility* boundaries,
3494 or crate boundaries (in order to simplify the module system and type checker).
3496 An example of a *recursive* type and its use:
3501 Cons(T, Box<List<T>>)
3504 let a: List<int> = Cons(7, box Cons(13, box Nil));
3509 All pointers in Rust are explicit first-class values.
3510 They can be copied, stored into data structures, and returned from functions.
3511 There are four varieties of pointer in Rust:
3513 * Owning pointers (`Box`)
3514 : These point to owned heap allocations (or "boxes") in the shared, inter-task heap.
3515 Each owned box has a single owning pointer; pointer and pointee retain a 1:1 relationship at all times.
3516 Owning pointers are written `Box<content>`,
3517 for example `Box<int>` means an owning pointer to an owned box containing an integer.
3518 Copying an owned box is a "deep" operation:
3519 it involves allocating a new owned box and copying the contents of the old box into the new box.
3520 Releasing an owning pointer immediately releases its corresponding owned box.
3523 : These point to memory _owned by some other value_.
3524 References arise by (automatic) conversion from owning pointers, managed pointers,
3525 or by applying the borrowing operator `&` to some other value,
3526 including [lvalues, rvalues or temporaries](#lvalues-rvalues-and-temporaries).
3527 References are written `&content`, or in some cases `&'f content` for some lifetime-variable `f`,
3528 for example `&int` means a reference to an integer.
3529 Copying a reference is a "shallow" operation:
3530 it involves only copying the pointer itself.
3531 Releasing a reference typically has no effect on the value it points to,
3532 with the exception of temporary values,
3533 which are released when the last reference to them is released.
3535 * Raw pointers (`*`)
3536 : Raw pointers are pointers without safety or liveness guarantees.
3537 Raw pointers are written `*content`,
3538 for example `*int` means a raw pointer to an integer.
3539 Copying or dropping a raw pointer has no effect on the lifecycle of any other value.
3540 Dereferencing a raw pointer or converting it to any other pointer type is an [`unsafe` operation](#unsafe-functions).
3541 Raw pointers are generally discouraged in Rust code;
3542 they exist to support interoperability with foreign code,
3543 and writing performance-critical or low-level functions.
3547 The function type constructor `fn` forms new function types.
3548 A function type consists of a possibly-empty set of function-type modifiers
3549 (such as `unsafe` or `extern`), a sequence of input types and an output type.
3551 An example of a `fn` type:
3554 fn add(x: int, y: int) -> int {
3558 let mut x = add(5,7);
3560 type Binop<'a> = |int,int|: 'a -> int;
3561 let bo: Binop = add;
3567 ~~~~ {.ebnf .notation}
3568 closure_type := [ 'unsafe' ] [ '<' lifetime-list '>' ] '|' arg-list '|'
3569 [ ':' bound-list ] [ '->' type ]
3570 procedure_type := 'proc' [ '<' lifetime-list '>' ] '(' arg-list ')'
3571 [ ':' bound-list ] [ '->' type ]
3572 lifetime-list := lifetime | lifetime ',' lifetime-list
3573 arg-list := ident ':' type | ident ':' type ',' arg-list
3574 bound-list := bound | bound '+' bound-list
3575 bound := path | lifetime
3578 The type of a closure mapping an input of type `A` to an output of type `B` is
3579 `|A| -> B`. A closure with no arguments or return values has type `||`.
3580 Similarly, a procedure mapping `A` to `B` is `proc(A) -> B` and a no-argument
3581 and no-return value closure has type `proc()`.
3583 An example of creating and calling a closure:
3586 let captured_var = 10;
3588 let closure_no_args = || println!("captured_var={}", captured_var);
3590 let closure_args = |arg: int| -> int {
3591 println!("captured_var={}, arg={}", captured_var, arg);
3592 arg // Note lack of semicolon after 'arg'
3595 fn call_closure(c1: ||, c2: |int| -> int) {
3600 call_closure(closure_no_args, closure_args);
3604 Unlike closures, procedures may only be invoked once, but own their
3605 environment, and are allowed to move out of their environment. Procedures are
3606 allocated on the heap (unlike closures). An example of creating and calling a
3610 let string = "Hello".to_string();
3612 // Creates a new procedure, passing it to the `spawn` function.
3614 println!("{} world!", string);
3617 // the variable `string` has been moved into the previous procedure, so it is
3618 // no longer usable.
3621 // Create an invoke a procedure. Note that the procedure is *moved* when
3622 // invoked, so it cannot be invoked again.
3623 let f = proc(n: int) { n + 22 };
3624 println!("answer: {}", f(20));
3630 Every trait item (see [traits](#traits)) defines a type with the same name as the trait.
3631 This type is called the _object type_ of the trait.
3632 Object types permit "late binding" of methods, dispatched using _virtual method tables_ ("vtables").
3633 Whereas most calls to trait methods are "early bound" (statically resolved) to specific implementations at compile time,
3634 a call to a method on an object type is only resolved to a vtable entry at compile time.
3635 The actual implementation for each vtable entry can vary on an object-by-object basis.
3637 Given a pointer-typed expression `E` of type `&T` or `Box<T>`, where `T` implements trait `R`,
3638 casting `E` to the corresponding pointer type `&R` or `Box<R>` results in a value of the _object type_ `R`.
3639 This result is represented as a pair of pointers:
3640 the vtable pointer for the `T` implementation of `R`, and the pointer value of `E`.
3642 An example of an object type:
3646 fn to_string(&self) -> String;
3649 impl Printable for int {
3650 fn to_string(&self) -> String { self.to_str() }
3653 fn print(a: Box<Printable>) {
3654 println!("{}", a.to_string());
3658 print(box 10 as Box<Printable>);
3662 In this example, the trait `Printable` occurs as an object type in both the type signature of `print`,
3663 and the cast expression in `main`.
3667 Within the body of an item that has type parameter declarations, the names of its type parameters are types:
3670 fn map<A: Clone, B: Clone>(f: |A| -> B, xs: &[A]) -> Vec<B> {
3674 let first: B = f(xs[0].clone());
3675 let rest: Vec<B> = map(f, xs.slice(1, xs.len()));
3676 return vec![first].append(rest.as_slice());
3680 Here, `first` has type `B`, referring to `map`'s `B` type parameter;
3681 and `rest` has type `Vec<B>`, a vector type with element type `B`.
3685 The special type `self` has a meaning within methods inside an
3686 impl item. It refers to the type of the implicit `self` argument. For
3691 fn make_string(&self) -> String;
3694 impl Printable for String {
3695 fn make_string(&self) -> String {
3701 `self` refers to the value of type `String` that is the receiver for a
3702 call to the method `make_string`.
3706 Types in Rust are categorized into kinds, based on various properties of the components of the type.
3710 : Types of this kind can be safely sent between tasks.
3711 This kind includes scalars, owning pointers, owned closures, and
3712 structural types containing only other owned types.
3713 All `Send` types are `'static`.
3715 : Types of this kind consist of "Plain Old Data"
3716 which can be copied by simply moving bits.
3717 All values of this kind can be implicitly copied.
3718 This kind includes scalars and immutable references,
3719 as well as structural types containing other `Copy` types.
3721 : Types of this kind do not contain any references (except for
3722 references with the `static` lifetime, which are allowed).
3723 This can be a useful guarantee for code
3724 that breaks borrowing assumptions
3725 using [`unsafe` operations](#unsafe-functions).
3727 : This is not strictly a kind,
3728 but its presence interacts with kinds:
3729 the `Drop` trait provides a single method `drop`
3730 that takes no parameters,
3731 and is run when values of the type are dropped.
3732 Such a method is called a "destructor",
3733 and are always executed in "top-down" order:
3734 a value is completely destroyed
3735 before any of the values it owns run their destructors.
3736 Only `Send` types can implement `Drop`.
3739 : Types with destructors, closure environments,
3740 and various other _non-first-class_ types,
3741 are not copyable at all.
3742 Such types can usually only be accessed through pointers,
3743 or in some cases, moved between mutable locations.
3745 Kinds can be supplied as _bounds_ on type parameters, like traits,
3746 in which case the parameter is constrained to types satisfying that kind.
3748 By default, type parameters do not carry any assumed kind-bounds at all.
3749 When instantiating a type parameter,
3750 the kind bounds on the parameter are checked
3751 to be the same or narrower than the kind
3752 of the type that it is instantiated with.
3754 Sending operations are not part of the Rust language,
3755 but are implemented in the library.
3756 Generic functions that send values
3757 bound the kind of these values to sendable.
3759 # Memory and concurrency models
3761 Rust has a memory model centered around concurrently-executing _tasks_. Thus
3762 its memory model and its concurrency model are best discussed simultaneously,
3763 as parts of each only make sense when considered from the perspective of the
3766 When reading about the memory model, keep in mind that it is partitioned in
3767 order to support tasks; and when reading about tasks, keep in mind that their
3768 isolation and communication mechanisms are only possible due to the ownership
3769 and lifetime semantics of the memory model.
3773 A Rust program's memory consists of a static set of *items*, a set of
3774 [tasks](#tasks) each with its own *stack*, and a *heap*. Immutable portions of
3775 the heap may be shared between tasks, mutable portions may not.
3777 Allocations in the stack consist of *slots*, and allocations in the heap
3780 ### Memory allocation and lifetime
3782 The _items_ of a program are those functions, modules and types
3783 that have their value calculated at compile-time and stored uniquely in the
3784 memory image of the rust process. Items are neither dynamically allocated nor
3787 A task's _stack_ consists of activation frames automatically allocated on
3788 entry to each function as the task executes. A stack allocation is reclaimed
3789 when control leaves the frame containing it.
3791 The _heap_ is a general term that describes two separate sets of boxes:
3792 managed boxes — which may be subject to garbage collection — and owned
3793 boxes. The lifetime of an allocation in the heap depends on the lifetime of
3794 the box values pointing to it. Since box values may themselves be passed in
3795 and out of frames, or stored in the heap, heap allocations may outlive the
3796 frame they are allocated within.
3798 ### Memory ownership
3800 A task owns all memory it can *safely* reach through local variables,
3801 as well as managed, owned boxes and references.
3803 When a task sends a value that has the `Send` trait to another task,
3804 it loses ownership of the value sent and can no longer refer to it.
3805 This is statically guaranteed by the combined use of "move semantics",
3806 and the compiler-checked _meaning_ of the `Send` trait:
3807 it is only instantiated for (transitively) sendable kinds of data constructor and pointers,
3808 never including managed boxes or references.
3810 When a stack frame is exited, its local allocations are all released, and its
3811 references to boxes (both managed and owned) are dropped.
3813 A managed box may (in the case of a recursive, mutable managed type) be cyclic;
3814 in this case the release of memory inside the managed structure may be deferred
3815 until task-local garbage collection can reclaim it. Code can ensure no such
3816 delayed deallocation occurs by restricting itself to owned boxes and similar
3817 unmanaged kinds of data.
3819 When a task finishes, its stack is necessarily empty and it therefore has no
3820 references to any boxes; the remainder of its heap is immediately freed.
3824 A task's stack contains slots.
3826 A _slot_ is a component of a stack frame, either a function parameter,
3827 a [temporary](#lvalues-rvalues-and-temporaries), or a local variable.
3829 A _local variable_ (or *stack-local* allocation) holds a value directly,
3830 allocated within the stack's memory. The value is a part of the stack frame.
3832 Local variables are immutable unless declared otherwise like: `let mut x = ...`.
3834 Function parameters are immutable unless declared with `mut`. The
3835 `mut` keyword applies only to the following parameter (so `|mut x, y|`
3836 and `fn f(mut x: Box<int>, y: Box<int>)` declare one mutable variable `x` and
3837 one immutable variable `y`).
3839 Methods that take either `self` or `~self` can optionally place them in a
3840 mutable slot by prefixing them with `mut` (similar to regular arguments):
3844 fn change(mut self) -> Self;
3845 fn modify(mut ~self) -> Box<Self>;
3849 Local variables are not initialized when allocated; the entire frame worth of
3850 local variables are allocated at once, on frame-entry, in an uninitialized
3851 state. Subsequent statements within a function may or may not initialize the
3852 local variables. Local variables can be used only after they have been
3853 initialized; this is enforced by the compiler.
3857 An _owned box_ is a reference to a heap allocation holding another value, which is constructed
3858 by the prefix operator `box`. When the standard library is in use, the type of an owned box is
3859 `std::owned::Box<T>`.
3861 An example of an owned box type and value:
3865 let x: Box<int> = box 10;
3868 Owned box values exist in 1:1 correspondence with their heap allocation
3869 copying an owned box value makes a shallow copy of the pointer
3870 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.
3873 let x: Box<int> = box 10;
3875 // attempting to use `x` will result in an error here
3882 An executing Rust program consists of a tree of tasks.
3883 A Rust _task_ consists of an entry function, a stack,
3884 a set of outgoing communication channels and incoming communication ports,
3885 and ownership of some portion of the heap of a single operating-system process.
3886 (We expect that many programs will not use channels and ports directly,
3887 but will instead use higher-level abstractions provided in standard libraries,
3890 Multiple Rust tasks may coexist in a single operating-system process.
3891 The runtime scheduler maps tasks to a certain number of operating-system threads.
3892 By default, the scheduler chooses the number of threads based on
3893 the number of concurrent physical CPUs detected at startup.
3894 It's also possible to override this choice at runtime.
3895 When the number of tasks exceeds the number of threads — which is likely —
3896 the scheduler multiplexes the tasks onto threads.[^mnscheduler]
3898 [^mnscheduler]: This is an M:N scheduler, which is known to give suboptimal
3899 results for CPU-bound concurrency problems. In such cases, running with the
3900 same number of threads and tasks can yield better results. Rust has M:N
3901 scheduling in order to support very large numbers of tasks in contexts where
3902 threads are too resource-intensive to use in large number. The cost of
3903 threads varies substantially per operating system, and is sometimes quite
3904 low, so this flexibility is not always worth exploiting.
3906 ### Communication between tasks
3908 Rust tasks are isolated and generally unable to interfere with one another's memory directly,
3909 except through [`unsafe` code](#unsafe-functions).
3910 All contact between tasks is mediated by safe forms of ownership transfer,
3911 and data races on memory are prohibited by the type system.
3913 Inter-task communication and co-ordination facilities are provided in the standard library.
3916 - synchronous and asynchronous communication channels with various communication topologies
3917 - read-only and read-write shared variables with various safe mutual exclusion patterns
3918 - simple locks and semaphores
3920 When such facilities carry values, the values are restricted to the [`Send` type-kind](#type-kinds).
3921 Restricting communication interfaces to this kind ensures that no references or managed pointers move between tasks.
3922 Thus access to an entire data structure can be mediated through its owning "root" value;
3923 no further locking or copying is required to avoid data races within the substructure of such a value.
3927 The _lifecycle_ of a task consists of a finite set of states and events
3928 that cause transitions between the states. The lifecycle states of a task are:
3935 A task begins its lifecycle — once it has been spawned — in the *running*
3936 state. In this state it executes the statements of its entry function, and any
3937 functions called by the entry function.
3939 A task may transition from the *running* state to the *blocked*
3940 state any time it makes a blocking communication call. When the
3941 call can be completed — when a message arrives at a sender, or a
3942 buffer opens to receive a message — then the blocked task will
3943 unblock and transition back to *running*.
3945 A task may transition to the *failing* state at any time, due being
3946 killed by some external event or internally, from the evaluation of a
3947 `fail!()` macro. Once *failing*, a task unwinds its stack and
3948 transitions to the *dead* state. Unwinding the stack of a task is done by
3949 the task itself, on its own control stack. If a value with a destructor is
3950 freed during unwinding, the code for the destructor is run, also on the task's
3951 control stack. Running the destructor code causes a temporary transition to a
3952 *running* state, and allows the destructor code to cause any subsequent
3953 state transitions. The original task of unwinding and failing thereby may
3954 suspend temporarily, and may involve (recursive) unwinding of the stack of a
3955 failed destructor. Nonetheless, the outermost unwinding activity will continue
3956 until the stack is unwound and the task transitions to the *dead*
3957 state. There is no way to "recover" from task failure. Once a task has
3958 temporarily suspended its unwinding in the *failing* state, failure
3959 occurring from within this destructor results in *hard* failure.
3960 A hard failure currently results in the process aborting.
3962 A task in the *dead* state cannot transition to other states; it exists
3963 only to have its termination status inspected by other tasks, and/or to await
3964 reclamation when the last reference to it drops.
3968 The currently scheduled task is given a finite *time slice* in which to
3969 execute, after which it is *descheduled* at a loop-edge or similar
3970 preemption point, and another task within is scheduled, pseudo-randomly.
3972 An executing task can yield control at any time, by making a library call to
3973 `std::task::yield`, which deschedules it immediately. Entering any other
3974 non-executing state (blocked, dead) similarly deschedules the task.
3976 # Runtime services, linkage and debugging
3978 The Rust _runtime_ is a relatively compact collection of C++ and Rust code
3979 that provides fundamental services and datatypes to all Rust tasks at
3980 run-time. It is smaller and simpler than many modern language runtimes. It is
3981 tightly integrated into the language's execution model of memory, tasks,
3982 communication and logging.
3984 > **Note:** The runtime library will merge with the `std` library in future versions of Rust.
3986 ### Memory allocation
3988 The runtime memory-management system is based on a _service-provider interface_,
3989 through which the runtime requests blocks of memory from its environment
3990 and releases them back to its environment when they are no longer needed.
3991 The default implementation of the service-provider interface
3992 consists of the C runtime functions `malloc` and `free`.
3994 The runtime memory-management system, in turn, supplies Rust tasks with
3995 facilities for allocating releasing stacks, as well as allocating and freeing
4000 The runtime provides C and Rust code to assist with various built-in types,
4001 such as vectors, strings, and the low level communication system (ports,
4004 Support for other built-in types such as simple types, tuples and
4005 enums is open-coded by the Rust compiler.
4007 ### Task scheduling and communication
4009 The runtime provides code to manage inter-task communication. This includes
4010 the system of task-lifecycle state transitions depending on the contents of
4011 queues, as well as code to copy values between queues and their recipients and
4012 to serialize values for transmission over operating-system inter-process
4013 communication facilities.
4017 The Rust compiler supports various methods to link crates together both
4018 statically and dynamically. This section will explore the various methods to
4019 link Rust crates together, and more information about native libraries can be
4020 found in the [ffi tutorial][ffi].
4022 In one session of compilation, the compiler can generate multiple artifacts
4023 through the usage of either command line flags or the `crate_type` attribute.
4024 If one or more command line flag is specified, all `crate_type` attributes will
4025 be ignored in favor of only building the artifacts specified by command line.
4027 * `--crate-type=bin`, `#[crate_type = "bin"]` - A runnable executable will be
4028 produced. This requires that there is a `main` function in the crate which
4029 will be run when the program begins executing. This will link in all Rust and
4030 native dependencies, producing a distributable binary.
4032 * `--crate-type=lib`, `#[crate_type = "lib"]` - A Rust library will be produced.
4033 This is an ambiguous concept as to what exactly is produced because a library
4034 can manifest itself in several forms. The purpose of this generic `lib` option
4035 is to generate the "compiler recommended" style of library. The output library
4036 will always be usable by rustc, but the actual type of library may change from
4037 time-to-time. The remaining output types are all different flavors of
4038 libraries, and the `lib` type can be seen as an alias for one of them (but the
4039 actual one is compiler-defined).
4041 * `--crate-type=dylib`, `#[crate_type = "dylib"]` - A dynamic Rust library will
4042 be produced. This is different from the `lib` output type in that this forces
4043 dynamic library generation. The resulting dynamic library can be used as a
4044 dependency for other libraries and/or executables. This output type will
4045 create `*.so` files on linux, `*.dylib` files on osx, and `*.dll` files on
4048 * `--crate-type=staticlib`, `#[crate_type = "staticlib"]` - A static system
4049 library will be produced. This is different from other library outputs in that
4050 the Rust compiler will never attempt to link to `staticlib` outputs. The
4051 purpose of this output type is to create a static library containing all of
4052 the local crate's code along with all upstream dependencies. The static
4053 library is actually a `*.a` archive on linux and osx and a `*.lib` file on
4054 windows. This format is recommended for use in situtations such as linking
4055 Rust code into an existing non-Rust application because it will not have
4056 dynamic dependencies on other Rust code.
4058 * `--crate-type=rlib`, `#[crate_type = "rlib"]` - A "Rust library" file will be
4059 produced. This is used as an intermediate artifact and can be thought of as a
4060 "static Rust library". These `rlib` files, unlike `staticlib` files, are
4061 interpreted by the Rust compiler in future linkage. This essentially means
4062 that `rustc` will look for metadata in `rlib` files like it looks for metadata
4063 in dynamic libraries. This form of output is used to produce statically linked
4064 executables as well as `staticlib` outputs.
4066 Note that these outputs are stackable in the sense that if multiple are
4067 specified, then the compiler will produce each form of output at once without
4068 having to recompile. However, this only applies for outputs specified by the same
4069 method. If only `crate_type` attributes are specified, then they will all be
4070 built, but if one or more `--crate-type` command line flag is specified,
4071 then only those outputs will be built.
4073 With all these different kinds of outputs, if crate A depends on crate B, then
4074 the compiler could find B in various different forms throughout the system. The
4075 only forms looked for by the compiler, however, are the `rlib` format and the
4076 dynamic library format. With these two options for a dependent library, the
4077 compiler must at some point make a choice between these two formats. With this
4078 in mind, the compiler follows these rules when determining what format of
4079 dependencies will be used:
4081 1. If a static library is being produced, all upstream dependencies are
4082 required to be available in `rlib` formats. This requirement stems from the
4083 reason that a dynamic library cannot be converted into a static format.
4085 Note that it is impossible to link in native dynamic dependencies to a static
4086 library, and in this case warnings will be printed about all unlinked native
4087 dynamic dependencies.
4089 2. If an `rlib` file is being produced, then there are no restrictions on what
4090 format the upstream dependencies are available in. It is simply required that
4091 all upstream dependencies be available for reading metadata from.
4093 The reason for this is that `rlib` files do not contain any of their upstream
4094 dependencies. It wouldn't be very efficient for all `rlib` files to contain a
4095 copy of `libstd.rlib`!
4097 3. If an executable is being produced and the `-C prefer-dynamic` flag is not
4098 specified, then dependencies are first attempted to be found in the `rlib`
4099 format. If some dependencies are not available in an rlib format, then
4100 dynamic linking is attempted (see below).
4102 4. If a dynamic library or an executable that is being dynamically linked is
4103 being produced, then the compiler will attempt to reconcile the available
4104 dependencies in either the rlib or dylib format to create a final product.
4106 A major goal of the compiler is to ensure that a library never appears more
4107 than once in any artifact. For example, if dynamic libraries B and C were
4108 each statically linked to library A, then a crate could not link to B and C
4109 together because there would be two copies of A. The compiler allows mixing
4110 the rlib and dylib formats, but this restriction must be satisfied.
4112 The compiler currently implements no method of hinting what format a library
4113 should be linked with. When dynamically linking, the compiler will attempt to
4114 maximize dynamic dependencies while still allowing some dependencies to be
4115 linked in via an rlib.
4117 For most situations, having all libraries available as a dylib is recommended
4118 if dynamically linking. For other situations, the compiler will emit a
4119 warning if it is unable to determine which formats to link each library with.
4121 In general, `--crate-type=bin` or `--crate-type=lib` should be sufficient for
4122 all compilation needs, and the other options are just available if more
4123 fine-grained control is desired over the output format of a Rust crate.
4127 The runtime contains a system for directing [logging
4128 expressions](#logging-expressions) to a logging console and/or internal logging
4129 buffers. Logging can be enabled per module.
4131 Logging output is enabled by setting the `RUST_LOG` environment
4132 variable. `RUST_LOG` accepts a logging specification made up of a
4133 comma-separated list of paths, with optional log levels. For each
4134 module containing log expressions, if `RUST_LOG` contains the path to
4135 that module or a parent of that module, then logs of the appropriate
4136 level will be output to the console.
4138 The path to a module consists of the crate name, any parent modules,
4139 then the module itself, all separated by double colons (`::`). The
4140 optional log level can be appended to the module path with an equals
4141 sign (`=`) followed by the log level, from 1 to 4, inclusive. Level 1
4142 is the error level, 2 is warning, 3 info, and 4 debug. You can also
4143 use the symbolic constants `error`, `warn`, `info`, and `debug`. Any
4144 logs less than or equal to the specified level will be output. If not
4145 specified then log level 4 is assumed. Debug messages can be omitted
4146 by passing `--cfg ndebug` to `rustc`.
4148 As an example, to see all the logs generated by the compiler, you would set
4149 `RUST_LOG` to `rustc`, which is the crate name (as specified in its `crate_id`
4150 [attribute](#attributes)). To narrow down the logs to just crate resolution,
4151 you would set it to `rustc::metadata::creader`. To see just error logging
4154 Note that when compiling source files that don't specify a
4155 crate name the crate is given a default name that matches the source file,
4156 with the extension removed. In that case, to turn on logging for a program
4157 compiled from, e.g. `helloworld.rs`, `RUST_LOG` should be set to `helloworld`.
4159 #### Logging Expressions
4161 Rust provides several macros to log information. Here's a simple Rust program
4162 that demonstrates all four of them:
4166 #[phase(plugin, link)] extern crate log;
4169 error!("This is an error log")
4170 warn!("This is a warn log")
4171 info!("this is an info log")
4172 debug!("This is a debug log")
4176 These four log levels correspond to levels 1-4, as controlled by `RUST_LOG`:
4179 $ RUST_LOG=rust=3 ./rust
4180 This is an error log
4185 # Appendix: Rationales and design tradeoffs
4189 # Appendix: Influences and further references
4193 > The essential problem that must be solved in making a fault-tolerant
4194 > software system is therefore that of fault-isolation. Different programmers
4195 > will write different modules, some modules will be correct, others will have
4196 > errors. We do not want the errors in one module to adversely affect the
4197 > behaviour of a module which does not have any errors.
4199 > — Joe Armstrong
4201 > In our approach, all data is private to some process, and processes can
4202 > only communicate through communications channels. *Security*, as used
4203 > in this paper, is the property which guarantees that processes in a system
4204 > cannot affect each other except by explicit communication.
4206 > When security is absent, nothing which can be proven about a single module
4207 > in isolation can be guaranteed to hold when that module is embedded in a
4210 > — Robert Strom and Shaula Yemini
4212 > Concurrent and applicative programming complement each other. The
4213 > ability to send messages on channels provides I/O without side effects,
4214 > while the avoidance of shared data helps keep concurrent processes from
4219 Rust is not a particularly original language. It may however appear unusual
4220 by contemporary standards, as its design elements are drawn from a number of
4221 "historical" languages that have, with a few exceptions, fallen out of
4222 favour. Five prominent lineages contribute the most, though their influences
4223 have come and gone during the course of Rust's development:
4225 * The NIL (1981) and Hermes (1990) family. These languages were developed by
4226 Robert Strom, Shaula Yemini, David Bacon and others in their group at IBM
4227 Watson Research Center (Yorktown Heights, NY, USA).
4229 * The Erlang (1987) language, developed by Joe Armstrong, Robert Virding, Claes
4230 Wikström, Mike Williams and others in their group at the Ericsson Computer
4231 Science Laboratory (Älvsjö, Stockholm, Sweden) .
4233 * The Sather (1990) language, developed by Stephen Omohundro, Chu-Cheow Lim,
4234 Heinz Schmidt and others in their group at The International Computer
4235 Science Institute of the University of California, Berkeley (Berkeley, CA,
4238 * The Newsqueak (1988), Alef (1995), and Limbo (1996) family. These
4239 languages were developed by Rob Pike, Phil Winterbottom, Sean Dorward and
4240 others in their group at Bell Labs Computing Sciences Research Center
4241 (Murray Hill, NJ, USA).
4243 * The Napier (1985) and Napier88 (1988) family. These languages were
4244 developed by Malcolm Atkinson, Ron Morrison and others in their group at
4245 the University of St. Andrews (St. Andrews, Fife, UK).
4247 Additional specific influences can be seen from the following languages:
4249 * The structural algebraic types and compilation manager of SML.
4250 * The attribute and assembly systems of C#.
4251 * The references and deterministic destructor system of C++.
4252 * The memory region systems of the ML Kit and Cyclone.
4253 * The typeclass system of Haskell.
4254 * The lexical identifier rule of Python.
4255 * The block syntax of Ruby.
4257 [ffi]: guide-ffi.html