1 % Rust Reference Manual
5 This document is the reference manual for the Rust programming language. It
6 provides three kinds of material:
8 - Chapters that formally define the language grammar and, for each
9 construct, informally describe its semantics and give examples of its
11 - Chapters that informally describe the memory model, concurrency model,
12 runtime services, linkage model and debugging facilities.
13 - Appendix chapters providing rationale and references to languages that
14 influenced the design.
16 This document does not serve as a tutorial introduction to the
17 language. Background familiarity with the language is assumed. A separate
18 [tutorial] document is available to help acquire such background familiarity.
20 This document also does not serve as a reference to the [standard] or [extra]
21 libraries included in the language distribution. Those libraries are
22 documented separately by extracting documentation attributes from their
25 [tutorial]: tutorial.html
26 [standard]: std/index.html
27 [extra]: extra/index.html
31 Rust is a work in progress. The language continues to evolve as the design
32 shifts and is fleshed out in working code. Certain parts work, certain parts
33 do not, certain parts will be removed or changed.
35 This manual is a snapshot written in the present tense. All features described
36 exist in working code unless otherwise noted, but some are quite primitive or
37 remain to be further modified by planned work. Some may be temporary. It is a
38 *draft*, and we ask that you not take anything you read here as final.
40 If you have suggestions to make, please try to focus them on *reductions* to
41 the language: possible features that can be combined or omitted. We aim to
42 keep the size and complexity of the language under control.
44 > **Note:** The grammar for Rust given in this document is rough and
45 > very incomplete; only a modest number of sections have accompanying grammar
46 > rules. Formalizing the grammar accepted by the Rust parser is ongoing work,
47 > but future versions of this document will contain a complete
48 > grammar. Moreover, we hope that this grammar will be extracted and verified
49 > as LL(1) by an automated grammar-analysis tool, and further tested against the
50 > Rust sources. Preliminary versions of this automation exist, but are not yet
55 Rust's grammar is defined over Unicode codepoints, each conventionally
56 denoted `U+XXXX`, for 4 or more hexadecimal digits `X`. _Most_ of Rust's
57 grammar is confined to the ASCII range of Unicode, and is described in this
58 document by a dialect of Extended Backus-Naur Form (EBNF), specifically a
59 dialect of EBNF supported by common automated LL(k) parsing tools such as
60 `llgen`, rather than the dialect given in ISO 14977. The dialect can be
61 defined self-referentially as follows:
63 ~~~~ {.ebnf .notation}
65 rule : nonterminal ':' productionrule ';' ;
66 productionrule : production [ '|' production ] * ;
68 term : element repeats ;
69 element : LITERAL | IDENTIFIER | '[' productionrule ']' ;
70 repeats : [ '*' | '+' ] NUMBER ? | NUMBER ? | '?' ;
75 - Whitespace in the grammar is ignored.
76 - Square brackets are used to group rules.
77 - `LITERAL` is a single printable ASCII character, or an escaped hexadecimal
78 ASCII code of the form `\xQQ`, in single quotes, denoting the corresponding
79 Unicode codepoint `U+00QQ`.
80 - `IDENTIFIER` is a nonempty string of ASCII letters and underscores.
81 - The `repeat` forms apply to the adjacent `element`, and are as follows:
82 - `?` means zero or one repetition
83 - `*` means zero or more repetitions
84 - `+` means one or more repetitions
85 - NUMBER trailing a repeat symbol gives a maximum repetition count
86 - NUMBER on its own gives an exact repetition count
88 This EBNF dialect should hopefully be familiar to many readers.
90 ## Unicode productions
92 A few productions in Rust's grammar permit Unicode codepoints outside the ASCII range.
93 We define these productions in terms of character properties specified in the Unicode standard,
94 rather than in terms of ASCII-range codepoints.
95 The section [Special Unicode Productions](#special-unicode-productions) lists these productions.
97 ## String table productions
99 Some rules in the grammar -- notably [unary
100 operators](#unary-operator-expressions), [binary
101 operators](#binary-operator-expressions), and [keywords](#keywords) --
102 are given in a simplified form: as a listing of a table of unquoted,
103 printable whitespace-separated strings. These cases form a subset of
104 the rules regarding the [token](#tokens) rule, and are assumed to be
105 the result of a lexical-analysis phase feeding the parser, driven by a
106 DFA, operating over the disjunction of all such string table entries.
108 When such a string enclosed in double-quotes (`"`) occurs inside the
109 grammar, it is an implicit reference to a single member of such a string table
110 production. See [tokens](#tokens) for more information.
116 Rust input is interpreted as a sequence of Unicode codepoints encoded in UTF-8,
117 normalized to Unicode normalization form NFKC.
118 Most Rust grammar rules are defined in terms of printable ASCII-range codepoints,
119 but a small number are defined in terms of Unicode properties or explicit codepoint lists.
120 ^[Substitute definitions for the special Unicode productions are provided to the grammar verifier, restricted to ASCII range, when verifying the grammar in this document.]
122 ## Special Unicode Productions
124 The following productions in the Rust grammar are defined in terms of Unicode properties:
125 `ident`, `non_null`, `non_star`, `non_eol`, `non_slash_or_star`, `non_single_quote` and `non_double_quote`.
129 The `ident` production is any nonempty Unicode string of the following form:
131 - The first character has property `XID_start`
132 - The remaining characters have property `XID_continue`
134 that does _not_ occur in the set of [keywords](#keywords).
136 Note: `XID_start` and `XID_continue` as character properties cover the
137 character ranges used to form the more familiar C and Java language-family
140 ### Delimiter-restricted productions
142 Some productions are defined by exclusion of particular Unicode characters:
144 - `non_null` is any single Unicode character aside from `U+0000` (null)
145 - `non_eol` is `non_null` restricted to exclude `U+000A` (`'\n'`)
146 - `non_star` is `non_null` restricted to exclude `U+002A` (`*`)
147 - `non_slash_or_star` is `non_null` restricted to exclude `U+002F` (`/`) and `U+002A` (`*`)
148 - `non_single_quote` is `non_null` restricted to exclude `U+0027` (`'`)
149 - `non_double_quote` is `non_null` restricted to exclude `U+0022` (`"`)
154 comment : block_comment | line_comment ;
155 block_comment : "/*" block_comment_body * '*' + '/' ;
156 block_comment_body : non_star * | '*' + non_slash_or_star ;
157 line_comment : "//" non_eol * ;
160 Comments in Rust code follow the general C++ style of line and block-comment forms,
161 with no nesting of block-comment delimiters.
163 Line comments beginning with _three_ slashes (`///`),
164 and block comments beginning with a repeated asterisk in the block-open sequence (`/**`),
165 are interpreted as a special syntax for `doc` [attributes](#attributes).
166 That is, they are equivalent to writing `#[doc "..."]` around the comment's text.
168 Non-doc comments are interpreted as a form of whitespace.
173 whitespace_char : '\x20' | '\x09' | '\x0a' | '\x0d' ;
174 whitespace : [ whitespace_char | comment ] + ;
177 The `whitespace_char` production is any nonempty Unicode string consisting of any
178 of the following Unicode characters: `U+0020` (space, `' '`), `U+0009` (tab,
179 `'\t'`), `U+000A` (LF, `'\n'`), `U+000D` (CR, `'\r'`).
181 Rust is a "free-form" language, meaning that all forms of whitespace serve
182 only to separate _tokens_ in the grammar, and have no semantic significance.
184 A Rust program has identical meaning if each whitespace element is replaced
185 with any other legal whitespace element, such as a single space character.
190 simple_token : keyword | unop | binop ;
191 token : simple_token | ident | literal | symbol | whitespace token ;
194 Tokens are primitive productions in the grammar defined by regular
195 (non-recursive) languages. "Simple" tokens are given in [string table
196 production](#string-table-productions) form, and occur in the rest of the
197 grammar as double-quoted strings. Other tokens have exact rules given.
201 The keywords are the following strings:
214 self static struct super
220 Each of these keywords has special meaning in its grammar,
221 and all of them are excluded from the `ident` rule.
225 A literal is an expression consisting of a single token, rather than a
226 sequence of tokens, that immediately and directly denotes the value it
227 evaluates to, rather than referring to it by name or some other evaluation
228 rule. A literal is a form of constant expression, so is evaluated (primarily)
232 literal : string_lit | char_lit | num_lit ;
235 #### Character and string literals
238 char_lit : '\x27' char_body '\x27' ;
239 string_lit : '"' string_body * '"' | 'r' raw_string ;
241 char_body : non_single_quote
242 | '\x5c' [ '\x27' | common_escape ] ;
244 string_body : non_double_quote
245 | '\x5c' [ '\x22' | common_escape ] ;
246 raw_string : '"' raw_string_body '"' | '#' raw_string '#' ;
248 common_escape : '\x5c'
249 | 'n' | 'r' | 't' | '0'
254 hex_digit : 'a' | 'b' | 'c' | 'd' | 'e' | 'f'
255 | 'A' | 'B' | 'C' | 'D' | 'E' | 'F'
257 oct_digit : '0' | '1' | '2' | '3' | '4' | '5' | '6' | '7' ;
258 dec_digit : '0' | nonzero_dec ;
259 nonzero_dec: '1' | '2' | '3' | '4'
260 | '5' | '6' | '7' | '8' | '9' ;
263 A _character literal_ is a single Unicode character enclosed within two
264 `U+0027` (single-quote) characters, with the exception of `U+0027` itself,
265 which must be _escaped_ by a preceding U+005C character (`\`).
267 A _string literal_ is a sequence of any Unicode characters enclosed within
268 two `U+0022` (double-quote) characters, with the exception of `U+0022`
269 itself, which must be _escaped_ by a preceding `U+005C` character (`\`),
270 or a _raw string literal_.
272 Some additional _escapes_ are available in either character or non-raw string
273 literals. An escape starts with a `U+005C` (`\`) and continues with one of
276 * An _8-bit codepoint escape_ escape starts with `U+0078` (`x`) and is
277 followed by exactly two _hex digits_. It denotes the Unicode codepoint
278 equal to the provided hex value.
279 * A _16-bit codepoint escape_ starts with `U+0075` (`u`) and is followed
280 by exactly four _hex digits_. It denotes the Unicode codepoint equal to
281 the provided hex value.
282 * A _32-bit codepoint escape_ starts with `U+0055` (`U`) and is followed
283 by exactly eight _hex digits_. It denotes the Unicode codepoint equal to
284 the provided hex value.
285 * A _whitespace escape_ is one of the characters `U+006E` (`n`), `U+0072`
286 (`r`), or `U+0074` (`t`), denoting the unicode values `U+000A` (LF),
287 `U+000D` (CR) or `U+0009` (HT) respectively.
288 * The _backslash escape_ is the character `U+005C` (`\`) which must be
289 escaped in order to denote *itself*.
291 Raw string literals do not process any escapes. They start with the character
292 `U+0072` (`r`), followed zero or more of the character `U+0023` (`#`) and a
293 `U+0022` (double-quote) character. The _raw string body_ is not defined in the
294 EBNF grammar above: it can contain any sequence of Unicode characters and is
295 terminated only by another `U+0022` (double-quote) character, followed by the
296 same number of `U+0023` (`#`) characters that preceeded the opening `U+0022`
297 (double-quote) character.
299 All Unicode characters contained in the raw string body represent themselves,
300 the characters `U+0022` (double-quote) (except when followed by at least as
301 many `U+0023` (`#`) characters as were used to start the raw string literal) or
302 `U+005C` (`\`) do not have any special meaning.
304 Examples for string literals:
307 "foo"; r"foo"; // foo
308 "\"foo\""; r#""foo""#; // "foo"
311 r##"foo #"# bar"##; // foo #"# bar
313 "\x52"; "R"; r"R"; // R
314 "\\x52"; r"\x52"; // \x52
321 num_lit : nonzero_dec [ dec_digit | '_' ] * num_suffix ?
322 | '0' [ [ dec_digit | '_' ] * num_suffix ?
323 | 'b' [ '1' | '0' | '_' ] + int_suffix ?
324 | 'o' [ oct_digit | '_' ] + int_suffix ?
325 | 'x' [ hex_digit | '_' ] + int_suffix ? ] ;
327 num_suffix : int_suffix | float_suffix ;
329 int_suffix : 'u' int_suffix_size ?
330 | 'i' int_suffix_size ? ;
331 int_suffix_size : [ '8' | '1' '6' | '3' '2' | '6' '4' ] ;
333 float_suffix : [ exponent | '.' dec_lit exponent ? ] ? float_suffix_ty ? ;
334 float_suffix_ty : 'f' [ '3' '2' | '6' '4' ] ;
335 exponent : ['E' | 'e'] ['-' | '+' ] ? dec_lit ;
336 dec_lit : [ dec_digit | '_' ] + ;
339 A _number literal_ is either an _integer literal_ or a _floating-point
340 literal_. The grammar for recognizing the two kinds of literals is mixed,
341 as they are differentiated by suffixes.
343 ##### Integer literals
345 An _integer literal_ has one of four forms:
347 * A _decimal literal_ starts with a *decimal digit* and continues with any
348 mixture of *decimal digits* and _underscores_.
349 * A _hex literal_ starts with the character sequence `U+0030` `U+0078`
350 (`0x`) and continues as any mixture hex digits and underscores.
351 * An _octal literal_ starts with the character sequence `U+0030` `U+006F`
352 (`0o`) and continues as any mixture octal digits and underscores.
353 * A _binary literal_ starts with the character sequence `U+0030` `U+0062`
354 (`0b`) and continues as any mixture binary digits and underscores.
356 An integer literal may be followed (immediately, without any spaces) by an
357 _integer suffix_, which changes the type of the literal. There are two kinds
358 of integer literal suffix:
360 * The `i` and `u` suffixes give the literal type `int` or `uint`,
362 * Each of the signed and unsigned machine types `u8`, `i8`,
363 `u16`, `i16`, `u32`, `i32`, `u64` and `i64`
364 give the literal the corresponding machine type.
366 The type of an _unsuffixed_ integer literal is determined by type inference.
367 If a integer type can be _uniquely_ determined from the surrounding program
368 context, the unsuffixed integer literal has that type. If the program context
369 underconstrains the type, the unsuffixed integer literal's type is `int`; if
370 the program context overconstrains the type, it is considered a static type
373 Examples of integer literals of various forms:
376 123; 0xff00; // type determined by program context
377 // defaults to int in absence of type
383 0o70_i16; // type i16
384 0b1111_1111_1001_0000_i32; // type i32
387 ##### Floating-point literals
389 A _floating-point literal_ has one of two forms:
391 * Two _decimal literals_ separated by a period
392 character `U+002E` (`.`), with an optional _exponent_ trailing after the
393 second decimal literal.
394 * A single _decimal literal_ followed by an _exponent_.
396 By default, a floating-point literal has a generic type, but will fall back to
397 `f64`. A floating-point literal may be followed (immediately, without any
398 spaces) by a _floating-point suffix_, which changes the type of the literal.
399 There are two floating-point suffixes: `f32`, and `f64` (the 32-bit and 64-bit
400 floating point types).
402 Examples of floating-point literals of various forms:
408 12E+99_f64; // type f64
411 ##### Unit and boolean literals
413 The _unit value_, the only value of the type that has the same name, is written as `()`.
414 The two values of the boolean type are written `true` and `false`.
420 | '#' | '[' | ']' | '(' | ')' | '{' | '}'
424 Symbols are a general class of printable [token](#tokens) that play structural
425 roles in a variety of grammar productions. They are catalogued here for
426 completeness as the set of remaining miscellaneous printable tokens that do not
427 otherwise appear as [unary operators](#unary-operator-expressions), [binary
428 operators](#binary-operator-expressions), or [keywords](#keywords).
434 expr_path : ident [ "::" expr_path_tail ] + ;
435 expr_path_tail : '<' type_expr [ ',' type_expr ] + '>'
438 type_path : ident [ type_path_tail ] + ;
439 type_path_tail : '<' type_expr [ ',' type_expr ] + '>'
443 A _path_ is a sequence of one or more path components _logically_ separated by
444 a namespace qualifier (`::`). If a path consists of only one component, it may
445 refer to either an [item](#items) or a [slot](#memory-slots) in a local
446 control scope. If a path has multiple components, it refers to an item.
448 Every item has a _canonical path_ within its crate, but the path naming an
449 item is only meaningful within a given crate. There is no global namespace
450 across crates; an item's canonical path merely identifies it within the crate.
452 Two examples of simple paths consisting of only identifier components:
459 Path components are usually [identifiers](#identifiers), but the trailing
460 component of a path may be an angle-bracket-enclosed list of type
461 arguments. In [expression](#expressions) context, the type argument list is
462 given after a final (`::`) namespace qualifier in order to disambiguate it
463 from a relational expression involving the less-than symbol (`<`). In type
464 expression context, the final namespace qualifier is omitted.
466 Two examples of paths with type arguments:
469 # use std::hashmap::HashMap;
471 # fn id<T>(t: T) -> T { t }
472 type t = HashMap<int,~str>; // Type arguments used in a type expression
473 let x = id::<int>(10); // Type arguments used in a call expression
479 A number of minor features of Rust are not central enough to have their own
480 syntax, and yet are not implementable as functions. Instead, they are given
481 names, and invoked through a consistent syntax: `name!(...)`. Examples
484 * `fmt!` : format data into a string
485 * `env!` : look up an environment variable's value at compile time
486 * `stringify!` : pretty-print the Rust expression given as an argument
487 * `proto!` : define a protocol for inter-task communication
488 * `include!` : include the Rust expression in the given file
489 * `include_str!` : include the contents of the given file as a string
490 * `include_bin!` : include the contents of the given file as a binary blob
491 * `error!`, `warn!`, `info!`, `debug!` : provide diagnostic information.
493 All of the above extensions, with the exception of `proto!`, are expressions
494 with values. `proto!` is an item, defining a new name.
499 expr_macro_rules : "macro_rules" '!' ident '(' macro_rule * ')'
500 macro_rule : '(' matcher * ')' "=>" '(' transcriber * ')' ';'
501 matcher : '(' matcher * ')' | '[' matcher * ']'
502 | '{' matcher * '}' | '$' ident ':' ident
503 | '$' '(' matcher * ')' sep_token? [ '*' | '+' ]
505 transcriber : '(' transcriber * ')' | '[' transcriber * ']'
506 | '{' transcriber * '}' | '$' ident
507 | '$' '(' transcriber * ')' sep_token? [ '*' | '+' ]
511 User-defined syntax extensions are called "macros",
512 and the `macro_rules` syntax extension defines them.
513 Currently, user-defined macros can expand to expressions, statements, or items.
515 (A `sep_token` is any token other than `*` and `+`.
516 A `non_special_token` is any token other than a delimiter or `$`.)
518 The macro expander looks up macro invocations by name,
519 and tries each macro rule in turn.
520 It transcribes the first successful match.
521 Matching and transcription are closely related to each other,
522 and we will describe them together.
526 The macro expander matches and transcribes every token that does not begin with a `$` literally, including delimiters.
527 For parsing reasons, delimiters must be balanced, but they are otherwise not special.
529 In the matcher, `$` _name_ `:` _designator_ matches the nonterminal in the
530 Rust syntax named by _designator_. Valid designators are `item`, `block`,
531 `stmt`, `pat`, `expr`, `ty` (type), `ident`, `path`, `matchers` (lhs of the `=>` in macro rules),
532 `tt` (rhs of the `=>` in macro rules). In the transcriber, the designator is already known, and so only
533 the name of a matched nonterminal comes after the dollar sign.
535 In both the matcher and transcriber, the Kleene star-like operator indicates repetition.
536 The Kleene star operator consists of `$` and parens, optionally followed by a separator token, followed by `*` or `+`.
537 `*` means zero or more repetitions, `+` means at least one repetition.
538 The parens are not matched or transcribed.
539 On the matcher side, a name is bound to _all_ of the names it
540 matches, in a structure that mimics the structure of the repetition
541 encountered on a successful match. The job of the transcriber is to sort that
544 The rules for transcription of these repetitions are called "Macro By Example".
545 Essentially, one "layer" of repetition is discharged at a time, and all of
546 them must be discharged by the time a name is transcribed. Therefore,
547 `( $( $i:ident ),* ) => ( $i )` is an invalid macro, but
548 `( $( $i:ident ),* ) => ( $( $i:ident ),* )` is acceptable (if trivial).
550 When Macro By Example encounters a repetition, it examines all of the `$`
551 _name_ s that occur in its body. At the "current layer", they all must repeat
552 the same number of times, so
553 ` ( $( $i:ident ),* ; $( $j:ident ),* ) => ( $( ($i,$j) ),* )` is valid if
554 given the argument `(a,b,c ; d,e,f)`, but not `(a,b,c ; d,e)`. The repetition
555 walks through the choices at that layer in lockstep, so the former input
556 transcribes to `( (a,d), (b,e), (c,f) )`.
558 Nested repetitions are allowed.
560 ### Parsing limitations
562 The parser used by the macro system is reasonably powerful, but the parsing of
563 Rust syntax is restricted in two ways:
565 1. The parser will always parse as much as possible. If it attempts to match
566 `$i:expr [ , ]` against `8 [ , ]`, it will attempt to parse `i` as an array
567 index operation and fail. Adding a separator can solve this problem.
568 2. The parser must have eliminated all ambiguity by the time it reaches a `$` _name_ `:` _designator_.
569 This requirement most often affects name-designator pairs when they occur at the beginning of, or immediately after, a `$(...)*`; requiring a distinctive token in front can solve the problem.
571 ## Syntax extensions useful for the macro author
573 * `log_syntax!` : print out the arguments at compile time
574 * `trace_macros!` : supply `true` or `false` to enable or disable macro expansion logging
575 * `stringify!` : turn the identifier argument into a string literal
576 * `concat!` : concatenates a comma-separated list of literals
577 * `concat_idents!` : create a new identifier by concatenating the arguments
579 # Crates and source files
581 Rust is a *compiled* language.
582 Its semantics obey a *phase distinction* between compile-time and run-time.
583 Those semantic rules that have a *static interpretation* govern the success or failure of compilation.
584 We refer to these rules as "static semantics".
585 Semantic rules called "dynamic semantics" govern the behavior of programs at run-time.
586 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.
588 The compilation model centres on artifacts called _crates_.
589 Each compilation processes a single crate in source form, and if successful, produces a single crate in binary form: either an executable or a library.^[A crate is somewhat
590 analogous to an *assembly* in the ECMA-335 CLI model, a *library* in the
591 SML/NJ Compilation Manager, a *unit* in the Owens and Flatt module system,
592 or a *configuration* in Mesa.]
594 A _crate_ is a unit of compilation and linking, as well as versioning, distribution and runtime loading.
595 A crate contains a _tree_ of nested [module](#modules) scopes.
596 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.
598 The Rust compiler is always invoked with a single source file as input, and always produces a single output crate.
599 The processing of that source file may result in other source files being loaded as modules.
600 Source files have the extension `.rs`.
602 A Rust source file describes a module, the name and
603 location of which -- in the module tree of the current crate -- are defined
604 from outside the source file: either by an explicit `mod_item` in
605 a referencing source file, or by the name of the crate itself.
607 Each source file contains a sequence of zero or more `item` definitions,
608 and may optionally begin with any number of `attributes` that apply to the containing module.
609 Attributes on the anonymous crate module define important metadata that influences
610 the behavior of the compiler.
613 // Linkage attributes
614 #[ link(name = "projx",
616 uuid = "9cccc5d5-aceb-4af5-8285-811211826b82") ];
618 // Additional metadata attributes
619 #[ desc = "Project X" ];
620 #[ license = "BSD" ];
621 #[ author = "Jane Doe" ];
623 // Specify the output type
624 #[ crate_type = "lib" ];
627 #[ warn(non_camel_case_types) ];
630 A crate that contains a `main` function can be compiled to an executable.
631 If a `main` function is present, its return type must be [`unit`](#primitive-types) and it must take no arguments.
633 # Items and attributes
635 Crates contain [items](#items),
636 each of which may have some number of [attributes](#attributes) attached to it.
641 item : mod_item | fn_item | type_item | struct_item | enum_item
642 | static_item | trait_item | impl_item | extern_block ;
645 An _item_ is a component of a crate; some module items can be defined in crate
646 files, but most are defined in source files. Items are organized within a
647 crate by a nested set of [modules](#modules). Every crate has a single
648 "outermost" anonymous module; all further items within the crate have
649 [paths](#paths) within the module tree of the crate.
651 Items are entirely determined at compile-time, generally remain fixed during
652 execution, and may reside in read-only memory.
654 There are several kinds of item:
656 * [modules](#modules)
657 * [functions](#functions)
658 * [type definitions](#type-definitions)
659 * [structures](#structures)
660 * [enumerations](#enumerations)
661 * [static items](#static-items)
663 * [implementations](#implementations)
665 Some items form an implicit scope for the declaration of sub-items. In other
666 words, within a function or module, declarations of items can (in many cases)
667 be mixed with the statements, control blocks, and similar artifacts that
668 otherwise compose the item body. The meaning of these scoped items is the same
669 as if the item was declared outside the scope -- it is still a static item --
670 except that the item's *path name* within the module namespace is qualified by
671 the name of the enclosing item, or is private to the enclosing item (in the
673 The grammar specifies the exact locations in which sub-item declarations may appear.
677 All items except modules may be *parameterized* by type. Type parameters are
678 given as a comma-separated list of identifiers enclosed in angle brackets
679 (`<...>`), after the name of the item and before its definition.
680 The type parameters of an item are considered "part of the name", not part of the type of the item.
681 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.
682 In practice, the type-inference system can usually infer such argument types from context.
683 There are no general type-parametric types, only type-parametric items.
684 That is, Rust has no notion of type abstraction: there are no first-class "forall" types.
689 mod_item : "mod" ident ( ';' | '{' mod '}' );
690 mod : [ view_item | item ] * ;
693 A module is a container for zero or more [view items](#view-items) and zero or
694 more [items](#items). The view items manage the visibility of the items
695 defined within the module, as well as the visibility of names from outside the
696 module when referenced from inside the module.
698 A _module item_ is a module, surrounded in braces, named, and prefixed with
699 the keyword `mod`. A module item introduces a new, named module into the tree
700 of modules making up a crate. Modules can nest arbitrarily.
702 An example of a module:
706 type complex = (f64, f64);
707 fn sin(f: f64) -> f64 {
711 fn cos(f: f64) -> f64 {
715 fn tan(f: f64) -> f64 {
722 Modules and types share the same namespace.
723 Declaring a named type that has the same name as a module in scope is forbidden:
724 that is, a type definition, trait, struct, enumeration, or type parameter
725 can't shadow the name of a module in scope, or vice versa.
727 A module without a body is loaded from an external file, by default with the same
728 name as the module, plus the `.rs` extension.
729 When a nested submodule is loaded from an external file,
730 it is loaded from a subdirectory path that mirrors the module hierarchy.
733 // Load the `vec` module from `vec.rs`
737 // Load the `local_data` module from `task/local_data.rs`
742 The directories and files used for loading external file modules can be influenced
743 with the `path` attribute.
746 #[path = "task_files"]
748 // Load the `local_data` module from `task_files/tls.rs`
757 view_item : extern_mod_decl | use_decl ;
760 A view item manages the namespace of a module.
761 View items do not define new items, but rather, simply change other items' visibility.
762 There are several kinds of view item:
764 * [`extern mod` declarations](#extern-mod-declarations)
765 * [`use` declarations](#use-declarations)
767 ##### Extern mod declarations
770 extern_mod_decl : "extern" "mod" ident [ '(' link_attrs ')' ] ? [ '=' string_lit ] ? ;
771 link_attrs : link_attr [ ',' link_attrs ] + ;
772 link_attr : ident '=' literal ;
775 An _`extern mod` declaration_ specifies a dependency on an external crate.
776 The external crate is then bound into the declaring scope
777 as the `ident` provided in the `extern_mod_decl`.
779 The external crate is resolved to a specific `soname` at compile time,
780 and a runtime linkage requirement to that `soname` is passed to the linker for
782 The `soname` is resolved at compile time by scanning the compiler's library path
783 and matching the `link_attrs` provided in the `use_decl` against any `#link` attributes that
784 were declared on the external crate when it was compiled.
785 If no `link_attrs` are provided,
786 a default `name` attribute is assumed,
787 equal to the `ident` given in the `use_decl`.
789 Optionally, an identifier in an `extern mod` declaration may be followed by an equals sign,
790 then a string literal denoting a relative path on the filesystem.
791 This path should exist in one of the directories in the Rust path,
792 which by default contains the `.rust` subdirectory of the current directory and each of its parents,
793 as well as any directories in the colon-separated (or semicolon-separated on Windows)
794 list of paths that is the `RUST_PATH` environment variable.
795 The meaning of `extern mod a = "b/c/d";`, supposing that `/a` is in the RUST_PATH,
796 is that the name `a` should be taken as a reference to the crate whose absolute location is
799 Four examples of `extern mod` declarations:
802 extern mod pcre (uuid = "54aba0f8-a7b1-4beb-92f1-4cf625264841");
804 extern mod extra; // equivalent to: extern mod extra ( name = "extra" );
806 extern mod rustextra (name = "extra"); // linking to 'extra' under another name
808 extern mod complicated_mod = "some-file/in/the-rust/path";
811 ##### Use declarations
814 use_decl : "pub" ? "use" ident [ '=' path
817 path_glob : ident [ "::" path_glob ] ?
819 | '{' ident [ ',' ident ] * '}'
822 A _use declaration_ creates one or more local name bindings synonymous
823 with some other [path](#paths).
824 Usually a `use` declaration is used to shorten the path required to refer to a module item.
826 *Note*: Unlike in many languages,
827 `use` declarations in Rust do *not* declare linkage dependency with external crates.
828 Rather, [`extern mod` declarations](#extern-mod-declarations) declare linkage dependencies.
830 Use declarations support a number of convenient shortcuts:
832 * Rebinding the target name as a new local name, using the syntax `use x = p::q::r;`.
833 * Simultaneously binding a list of paths differing only in their final element,
834 using the glob-like brace syntax `use a::b::{c,d,e,f};`
835 * Binding all paths matching a given prefix, using the asterisk wildcard syntax `use a::b::*;`
837 An example of `use` declarations:
841 use std::option::{Some, None};
846 // Equivalent to 'std::num::sin(1.0);'
849 // Equivalent to 'foo(~[std::option::Some(1.0), std::option::None]);'
850 foo(~[Some(1.0), None]);
854 Like items, `use` declarations are private to the containing module, by default.
855 Also like items, a `use` declaration can be public, if qualified by the `pub` keyword.
856 Such a `use` declaration serves to _re-export_ a name.
857 A public `use` declaration can therefore _redirect_ some public name to a different target definition:
858 even a definition with a private canonical path, inside a different module.
859 If a sequence of such redirections form a cycle or cannot be resolved unambiguously,
860 they represent a compile-time error.
862 An example of re-exporting:
866 pub use quux::foo::*;
875 In this example, the module `quux` re-exports all of the public names defined in `foo`.
877 Also note that the paths contained in `use` items are relative to the crate root.
878 So, in the previous example, the `use` refers to `quux::foo::*`, and not simply to `foo::*`.
879 This also means that top-level module declarations should be at the crate root if direct usage
880 of the declared modules within `use` items is desired. It is also possible to use `self` and `super`
881 at the beginning of a `use` item to refer to the current and direct parent modules respectively.
882 All rules regarding accessing declared modules in `use` declarations applies to both module declarations
883 and `extern mod` declarations.
885 An example of what will and will not work for `use` items:
887 # #[allow(unused_imports)];
888 use foo::extra; // good: foo is at the root of the crate
889 use foo::baz::foobaz; // good: foo is at the root of the crate
894 use foo::extra::list; // good: foo is at crate root
895 // use extra::*; // bad: extra is not at the crate root
896 use self::baz::foobaz; // good: self refers to module 'foo'
897 use foo::bar::foobar; // good: foo is at crate root
904 use super::bar::foobar; // good: super refers to module 'foo'
914 A _function item_ defines a sequence of [statements](#statements) and an optional final [expression](#expressions), along with a name and a set of parameters.
915 Functions are declared with the keyword `fn`.
916 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.
918 A function may also be copied into a first class *value*, in which case the
919 value has the corresponding [*function type*](#function-types), and can be
920 used otherwise exactly as a function item (with a minor additional cost of
921 calling the function indirectly).
923 Every control path in a function logically ends with a `return` expression or a
924 diverging expression. If the outermost block of a function has a
925 value-producing expression in its final-expression position, that expression
926 is interpreted as an implicit `return` expression applied to the
929 An example of a function:
932 fn add(x: int, y: int) -> int {
937 As with `let` bindings, function arguments are irrefutable patterns,
938 so any pattern that is valid in a let binding is also valid as an argument.
941 fn first((value, _): (int, int)) -> int { value }
945 #### Generic functions
947 A _generic function_ allows one or more _parameterized types_ to
948 appear in its signature. Each type parameter must be explicitly
949 declared, in an angle-bracket-enclosed, comma-separated list following
953 fn iter<T>(seq: &[T], f: |T|) {
954 for elt in seq.iter() { f(elt); }
956 fn map<T, U>(seq: &[T], f: |T| -> U) -> ~[U] {
958 for elt in seq.iter() { acc.push(f(elt)); }
963 Inside the function signature and body, the name of the type parameter
964 can be used as a type name.
966 When a generic function is referenced, its type is instantiated based
967 on the context of the reference. For example, calling the `iter`
968 function defined above on `[1, 2]` will instantiate type parameter `T`
969 with `int`, and require the closure parameter to have type
972 The type parameters can also be explicitly supplied in a trailing
973 [path](#paths) component after the function name. This might be necessary
974 if there is not sufficient context to determine the type parameters. For
975 example, `mem::size_of::<u32>() == 4`.
977 Since a parameter type is opaque to the generic function, the set of
978 operations that can be performed on it is limited. Values of parameter
979 type can only be moved, not copied.
982 fn id<T>(x: T) -> T { x }
985 Similarly, [trait](#traits) bounds can be specified for type
986 parameters to allow methods with that trait to be called on values
992 Unsafe operations are those that potentially violate the memory-safety guarantees of Rust's static semantics.
994 The following language level features cannot be used in the safe subset of Rust:
996 - Dereferencing a [raw pointer](#pointer-types).
997 - Calling an unsafe function (including an intrinsic or foreign function).
999 ##### Unsafe functions
1001 Unsafe functions are functions that are not safe in all contexts and/or for all possible inputs.
1002 Such a function must be prefixed with the keyword `unsafe`.
1006 A block of code can also be prefixed with the `unsafe` keyword, to permit calling `unsafe` functions
1007 or dereferencing raw pointers within a safe function.
1009 When a programmer has sufficient conviction that a sequence of potentially unsafe operations is
1010 actually safe, they can encapsulate that sequence (taken as a whole) within an `unsafe` block. The
1011 compiler will consider uses of such code safe, in the surrounding context.
1013 Unsafe blocks are used to wrap foreign libraries, make direct use of hardware or implement features
1014 not directly present in the language. For example, Rust provides the language features necessary to
1015 implement memory-safe concurrency in the language but the implementation of tasks and message
1016 passing is in the standard library.
1018 Rust's type system is a conservative approximation of the dynamic safety requirements, so in some
1019 cases there is a performance cost to using safe code. For example, a doubly-linked list is not a
1020 tree structure and can only be represented with managed or reference-counted pointers in safe code.
1021 By using `unsafe` blocks to represent the reverse links as raw pointers, it can be implemented with
1022 only owned pointers.
1024 ##### Behavior considered unsafe
1026 This is a list of behavior which is forbidden in all Rust code. Type checking provides the guarantee
1027 that these issues are never caused by safe code. An `unsafe` block or function is responsible for
1028 never invoking this behaviour or exposing an API making it possible for it to occur in safe code.
1031 * Dereferencing a null/dangling raw pointer
1032 * Mutating an immutable value/reference, if it is not marked as non-`Freeze`
1033 * Reads of [undef](http://llvm.org/docs/LangRef.html#undefined-values) (uninitialized) memory
1034 * Breaking the [pointer aliasing rules](http://llvm.org/docs/LangRef.html#pointer-aliasing-rules)
1035 with raw pointers (a subset of the rules used by C)
1036 * Invoking undefined behavior via compiler intrinsics:
1037 * Indexing outside of the bounds of an object with `std::ptr::offset` (`offset` intrinsic), with
1038 the exception of one byte past the end which is permitted.
1039 * Using `std::ptr::copy_nonoverlapping_memory` (`memcpy32`/`memcpy64` instrinsics) on
1041 * Invalid values in primitive types, even in private fields/locals:
1042 * Dangling/null pointers in non-raw pointers, or slices
1043 * A value other than `false` (0) or `true` (1) in a `bool`
1044 * A discriminant in an `enum` not included in the type definition
1045 * A value in a `char` which is a surrogate or above `char::MAX`
1046 * non-UTF-8 byte sequences in a `str`
1048 ##### Behaviour not considered unsafe
1050 This is a list of behaviour not considered *unsafe* in Rust terms, but that may be undesired.
1053 * Reading data from private fields (`std::repr`, `format!("{:?}", x)`)
1054 * Leaks due to reference count cycles, even in the global heap
1055 * Exiting without calling destructors
1057 * Accessing/modifying the file system
1058 * Unsigned integer overflow (well-defined as wrapping)
1059 * Signed integer overflow (well-defined as two's complement representation wrapping)
1061 #### Diverging functions
1063 A special kind of function can be declared with a `!` character where the
1064 output slot type would normally be. For example:
1067 fn my_err(s: &str) -> ! {
1073 We call such functions "diverging" because they never return a value to the
1074 caller. Every control path in a diverging function must end with a
1075 `fail!()` or a call to another diverging function on every
1076 control path. The `!` annotation does *not* denote a type. Rather, the result
1077 type of a diverging function is a special type called $\bot$ ("bottom") that
1078 unifies with any type. Rust has no syntax for $\bot$.
1080 It might be necessary to declare a diverging function because as mentioned
1081 previously, the typechecker checks that every control path in a function ends
1082 with a [`return`](#return-expressions) or diverging expression. So, if `my_err`
1083 were declared without the `!` annotation, the following code would not
1087 # fn my_err(s: &str) -> ! { fail!() }
1089 fn f(i: int) -> int {
1094 my_err("Bad number!");
1099 This will not compile without the `!` annotation on `my_err`,
1100 since the `else` branch of the conditional in `f` does not return an `int`,
1101 as required by the signature of `f`.
1102 Adding the `!` annotation to `my_err` informs the typechecker that,
1103 should control ever enter `my_err`, no further type judgments about `f` need to hold,
1104 since control will never resume in any context that relies on those judgments.
1105 Thus the return type on `f` only needs to reflect the `if` branch of the conditional.
1108 #### Extern functions
1110 Extern functions are part of Rust's foreign function interface,
1111 providing the opposite functionality to [external blocks](#external-blocks).
1112 Whereas external blocks allow Rust code to call foreign code,
1113 extern functions with bodies defined in Rust code _can be called by foreign
1114 code_. They are defined in the same way as any other Rust function,
1115 except that they have the `extern` modifier.
1118 // Declares an extern fn, the ABI defaults to "C"
1119 extern fn new_vec() -> ~[int] { ~[] }
1121 // Declares an extern fn with "stdcall" ABI
1122 extern "stdcall" fn new_vec_stdcall() -> ~[int] { ~[] }
1125 Unlike normal functions, extern fns have an `extern "ABI" fn()`.
1126 This is the same type as the functions declared in an extern
1130 # extern fn new_vec() -> ~[int] { ~[] }
1131 let fptr: extern "C" fn() -> ~[int] = new_vec;
1134 Extern functions may be called directly from Rust code as Rust uses large,
1135 contiguous stack segments like C.
1137 ### Type definitions
1139 A _type definition_ defines a new name for an existing [type](#types). Type
1140 definitions are declared with the keyword `type`. Every value has a single,
1141 specific type; the type-specified aspects of a value include:
1143 * Whether the value is composed of sub-values or is indivisible.
1144 * Whether the value represents textual or numerical information.
1145 * Whether the value represents integral or floating-point information.
1146 * The sequence of memory operations required to access the value.
1147 * The [kind](#type-kinds) of the type.
1149 For example, the type `(u8, u8)` defines the set of immutable values that are composite pairs,
1150 each containing two unsigned 8-bit integers accessed by pattern-matching and laid out in memory with the `x` component preceding the `y` component.
1154 A _structure_ is a nominal [structure type](#structure-types) defined with the keyword `struct`.
1156 An example of a `struct` item and its use:
1159 struct Point {x: int, y: int}
1160 let p = Point {x: 10, y: 11};
1164 A _tuple structure_ is a nominal [tuple type](#tuple-types), also defined with the keyword `struct`.
1168 struct Point(int, int);
1169 let p = Point(10, 11);
1170 let px: int = match p { Point(x, _) => x };
1173 A _unit-like struct_ is a structure without any fields, defined by leaving off the list of fields entirely.
1174 Such types will have a single value, just like the [unit value `()`](#unit-and-boolean-literals) of the unit type.
1179 let c = [Cookie, Cookie, Cookie, Cookie];
1184 An _enumeration_ is a simultaneous definition of a nominal [enumerated type](#enumerated-types) as well as a set of *constructors*,
1185 that can be used to create or pattern-match values of the corresponding enumerated type.
1187 Enumerations are declared with the keyword `enum`.
1189 An example of an `enum` item and its use:
1197 let mut a: Animal = Dog;
1201 Enumeration constructors can have either named or unnamed fields:
1205 Cat { name: ~str, weight: f64 }
1208 let mut a: Animal = Dog(~"Cocoa", 37.2);
1209 a = Cat{ name: ~"Spotty", weight: 2.7 };
1212 In this example, `Cat` is a _struct-like enum variant_,
1213 whereas `Dog` is simply called an enum variant.
1218 static_item : "static" ident ':' type '=' expr ';' ;
1221 A *static item* is a named _constant value_ stored in the global data section of a crate.
1222 Immutable static items are stored in the read-only data section.
1223 The constant value bound to a static item is, like all constant values, evaluated at compile time.
1224 Static items have the `static` lifetime, which outlives all other lifetimes in a Rust program.
1225 Static items are declared with the `static` keyword.
1226 A static item must have a _constant expression_ giving its definition.
1228 Static items must be explicitly typed.
1229 The type may be ```bool```, ```char```, a number, or a type derived from those primitive types.
1230 The derived types are borrowed pointers with the `'static` lifetime,
1231 fixed-size arrays, tuples, and structs.
1234 static BIT1: uint = 1 << 0;
1235 static BIT2: uint = 1 << 1;
1237 static BITS: [uint, ..2] = [BIT1, BIT2];
1238 static STRING: &'static str = "bitstring";
1240 struct BitsNStrings<'self> {
1241 mybits: [uint, ..2],
1242 mystring: &'self str
1245 static bits_n_strings: BitsNStrings<'static> = BitsNStrings {
1251 #### Mutable statics
1253 If a static item is declared with the ```mut``` keyword, then it is allowed to
1254 be modified by the program. One of Rust's goals is to make concurrency bugs hard
1255 to run into, and this is obviously a very large source of race conditions or
1256 other bugs. For this reason, an ```unsafe``` block is required when either
1257 reading or writing a mutable static variable. Care should be taken to ensure
1258 that modifications to a mutable static are safe with respect to other tasks
1259 running in the same process.
1261 Mutable statics are still very useful, however. They can be used with C
1262 libraries and can also be bound from C libraries (in an ```extern``` block).
1265 # fn atomic_add(_: &mut uint, _: uint) -> uint { 2 }
1267 static mut LEVELS: uint = 0;
1269 // This violates the idea of no shared state, and this doesn't internally
1270 // protect against races, so this function is `unsafe`
1271 unsafe fn bump_levels_unsafe1() -> uint {
1277 // Assuming that we have an atomic_add function which returns the old value,
1278 // this function is "safe" but the meaning of the return value may not be what
1279 // callers expect, so it's still marked as `unsafe`
1280 unsafe fn bump_levels_unsafe2() -> uint {
1281 return atomic_add(&mut LEVELS, 1);
1287 A _trait_ describes a set of method types.
1289 Traits can include default implementations of methods,
1290 written in terms of some unknown [`self` type](#self-types);
1291 the `self` type may either be completely unspecified,
1292 or constrained by some other trait.
1294 Traits are implemented for specific types through separate [implementations](#implementations).
1297 # type Surface = int;
1298 # type BoundingBox = int;
1301 fn draw(&self, Surface);
1302 fn bounding_box(&self) -> BoundingBox;
1306 This defines a trait with two methods.
1307 All values that have [implementations](#implementations) of this trait in scope can have their `draw` and `bounding_box` methods called,
1308 using `value.bounding_box()` [syntax](#method-call-expressions).
1310 Type parameters can be specified for a trait to make it generic.
1311 These appear after the trait name, using the same syntax used in [generic functions](#generic-functions).
1315 fn len(&self) -> uint;
1316 fn elt_at(&self, n: uint) -> T;
1317 fn iter(&self, |T|);
1321 Generic functions may use traits as _bounds_ on their type parameters.
1322 This will have two effects: only types that have the trait may instantiate the parameter,
1323 and within the generic function,
1324 the methods of the trait can be called on values that have the parameter's type.
1328 # type Surface = int;
1329 # trait Shape { fn draw(&self, Surface); }
1331 fn draw_twice<T: Shape>(surface: Surface, sh: T) {
1337 Traits also define an [object type](#object-types) with the same name as the trait.
1338 Values of this type are created by [casting](#type-cast-expressions) pointer values
1339 (pointing to a type for which an implementation of the given trait is in scope)
1340 to pointers to the trait name, used as a type.
1344 # impl Shape for int { }
1347 let myshape: @Shape = @mycircle as @Shape;
1350 The resulting value is a managed box containing the value that was cast,
1351 along with information that identifies the methods of the implementation that was used.
1352 Values with a trait type can have [methods called](#method-call-expressions) on them,
1353 for any method in the trait,
1354 and can be used to instantiate type parameters that are bounded by the trait.
1356 Trait methods may be static,
1357 which means that they lack a `self` argument.
1358 This means that they can only be called with function call syntax (`f(x)`)
1359 and not method call syntax (`obj.f()`).
1360 The way to refer to the name of a static method is to qualify it with the trait name,
1361 treating the trait name like a module.
1366 fn from_int(n: int) -> Self;
1369 fn from_int(n: int) -> f64 { n as f64 }
1371 let x: f64 = Num::from_int(42);
1374 Traits may inherit from other traits. For example, in
1377 trait Shape { fn area() -> f64; }
1378 trait Circle : Shape { fn radius() -> f64; }
1381 the syntax `Circle : Shape` means that types that implement `Circle` must also have an implementation for `Shape`.
1382 Multiple supertraits are separated by spaces, `trait Circle : Shape Eq { }`.
1383 In an implementation of `Circle` for a given type `T`, methods can refer to `Shape` methods,
1384 since the typechecker checks that any type with an implementation of `Circle` also has an implementation of `Shape`.
1386 In type-parameterized functions,
1387 methods of the supertrait may be called on values of subtrait-bound type parameters.
1388 Referring to the previous example of `trait Circle : Shape`:
1391 # trait Shape { fn area(&self) -> f64; }
1392 # trait Circle : Shape { fn radius(&self) -> f64; }
1393 fn radius_times_area<T: Circle>(c: T) -> f64 {
1394 // `c` is both a Circle and a Shape
1395 c.radius() * c.area()
1399 Likewise, supertrait methods may also be called on trait objects.
1402 # trait Shape { fn area(&self) -> f64; }
1403 # trait Circle : Shape { fn radius(&self) -> f64; }
1404 # impl Shape for int { fn area(&self) -> f64 { 0.0 } }
1405 # impl Circle for int { fn radius(&self) -> f64 { 0.0 } }
1408 let mycircle: Circle = @mycircle as @Circle;
1409 let nonsense = mycircle.radius() * mycircle.area();
1414 An _implementation_ is an item that implements a [trait](#traits) for a specific type.
1416 Implementations are defined with the keyword `impl`.
1419 # struct Point {x: f64, y: f64};
1420 # type Surface = int;
1421 # struct BoundingBox {x: f64, y: f64, width: f64, height: f64};
1422 # trait Shape { fn draw(&self, Surface); fn bounding_box(&self) -> BoundingBox; }
1423 # fn do_draw_circle(s: Surface, c: Circle) { }
1430 impl Shape for Circle {
1431 fn draw(&self, s: Surface) { do_draw_circle(s, *self); }
1432 fn bounding_box(&self) -> BoundingBox {
1433 let r = self.radius;
1434 BoundingBox{x: self.center.x - r, y: self.center.y - r,
1435 width: 2.0 * r, height: 2.0 * r}
1440 It is possible to define an implementation without referring to a trait.
1441 The methods in such an implementation can only be used
1442 as direct calls on the values of the type that the implementation targets.
1443 In such an implementation, the trait type and `for` after `impl` are omitted.
1444 Such implementations are limited to nominal types (enums, structs),
1445 and the implementation must appear in the same module or a sub-module as the `self` type.
1447 When a trait _is_ specified in an `impl`,
1448 all methods declared as part of the trait must be implemented,
1449 with matching types and type parameter counts.
1451 An implementation can take type parameters,
1452 which can be different from the type parameters taken by the trait it implements.
1453 Implementation parameters are written after the `impl` keyword.
1458 impl<T> Seq<T> for ~[T] {
1461 impl Seq<bool> for u32 {
1462 /* Treat the integer as a sequence of bits */
1469 extern_block_item : "extern" '{' extern_block '} ;
1470 extern_block : [ foreign_fn ] * ;
1473 External blocks form the basis for Rust's foreign function interface.
1474 Declarations in an external block describe symbols
1475 in external, non-Rust libraries.
1477 Functions within external blocks
1478 are declared in the same way as other Rust functions,
1479 with the exception that they may not have a body
1480 and are instead terminated by a semicolon.
1483 # use std::libc::{c_char, FILE};
1487 fn fopen(filename: *c_char, mode: *c_char) -> *FILE;
1491 Functions within external blocks may be called by Rust code,
1492 just like functions defined in Rust.
1493 The Rust compiler automatically translates
1494 between the Rust ABI and the foreign ABI.
1496 A number of [attributes](#attributes) control the behavior of external
1499 By default external blocks assume that the library they are calling
1500 uses the standard C "cdecl" ABI. Other ABIs may be specified using
1501 an `abi` string, as shown here:
1504 // Interface to the Windows API
1505 extern "stdcall" { }
1508 The `link_name` attribute allows the name of the library to be specified.
1511 #[link_name = "crypto"]
1515 The `nolink` attribute tells the Rust compiler
1516 not to do any linking for the external block.
1517 This is particularly useful for creating external blocks for libc,
1518 which tends to not follow standard library naming conventions
1519 and is linked to all Rust programs anyway.
1521 The type of a function
1522 declared in an extern block
1523 is `extern "abi" fn(A1, ..., An) -> R`,
1524 where `A1...An` are the declared types of its arguments
1525 and `R` is the decalred return type.
1527 ## Visibility and Privacy
1529 These two terms are often used interchangeably, and what they are attempting to
1530 convey is the answer to the question "Can this item be used at this location?"
1532 Rust's name resolution operates on a global hierarchy of namespaces. Each level
1533 in the hierarchy can be thought of as some item. The items are one of those
1534 mentioned above, but also include external crates. Declaring or defining a new
1535 module can be thought of as inserting a new tree into the hierarchy at the
1536 location of the definition.
1538 To control whether interfaces can be used across modules, Rust checks each use
1539 of an item to see whether it should be allowed or not. This is where privacy
1540 warnings are generated, or otherwise "you used a private item of another module
1541 and weren't allowed to."
1543 By default, everything in rust is *private*, with two exceptions. The first
1544 exception is that struct fields are public by default (but the struct itself is
1545 still private by default), and the remaining exception is that enum variants in
1546 a `pub` enum are the default visibility of the enum container itself.. You are
1547 allowed to alter this default visibility with the `pub` keyword (or `priv`
1548 keyword for struct fields and enum variants). When an item is declared as `pub`,
1549 it can be thought of as being accessible to the outside world. For example:
1553 // Declare a private struct
1556 // Declare a public struct with a private field
1561 // Declare a public enum with public and private variants
1563 PubliclyAccessibleState,
1564 priv PrivatelyAccessibleState
1568 With the notion of an item being either public or private, Rust allows item
1569 accesses in two cases:
1571 1. If an item is public, then it can be used externally through any of its
1573 2. If an item is private, it may be accessed by the current module and its
1576 These two cases are surprisingly powerful for creating module hierarchies
1577 exposing public APIs while hiding internal implementation details. To help
1578 explain, here's a few use cases and what they would entail.
1580 * A library developer needs to expose functionality to crates which link against
1581 their library. As a consequence of the first case, this means that anything
1582 which is usable externally must be `pub` from the root down to the destination
1583 item. Any private item in the chain will disallow external accesses.
1585 * A crate needs a global available "helper module" to itself, but it doesn't
1586 want to expose the helper module as a public API. To accomplish this, the root
1587 of the crate's hierarchy would have a private module which then internally has
1588 a "public api". Because the entire crate is a descendant of the root, then the
1589 entire local crate can access this private module through the second case.
1591 * When writing unit tests for a module, it's often a common idiom to have an
1592 immediate child of the module to-be-tested named `mod test`. This module could
1593 access any items of the parent module through the second case, meaning that
1594 internal implementation details could also be seamlessly tested from the child
1597 In the second case, it mentions that a private item "can be accessed" by the
1598 current module and its descendants, but the exact meaning of accessing an item
1599 depends on what the item is. Accessing a module, for example, would mean looking
1600 inside of it (to import more items). On the other hand, accessing a function
1601 would mean that it is invoked. Additionally, path expressions and import
1602 statements are considered to access an item in the sense that the
1603 import/expression is only valid if the destination is in the current visibility
1606 Here's an example of a program which exemplifies the three cases outlined above.
1609 // This module is private, meaning that no external crate can access this
1610 // module. Because it is private at the root of this current crate, however, any
1611 // module in the crate may access any publicly visible item in this module.
1612 mod crate_helper_module {
1614 // This function can be used by anything in the current crate
1615 pub fn crate_helper() {}
1617 // This function *cannot* be used by anything else in the crate. It is not
1618 // publicly visible outside of the `crate_helper_module`, so only this
1619 // current module and its descendants may access it.
1620 fn implementation_detail() {}
1623 // This function is "public to the root" meaning that it's available to external
1624 // crates linking against this one.
1625 pub fn public_api() {}
1627 // Similarly to 'public_api', this module is public so external crates may look
1630 use crate_helper_module;
1632 pub fn my_method() {
1633 // Any item in the local crate may invoke the helper module's public
1634 // interface through a combination of the two rules above.
1635 crate_helper_module::crate_helper();
1638 // This function is hidden to any module which is not a descendant of
1640 fn my_implementation() {}
1646 fn test_my_implementation() {
1647 // Because this module is a descendant of `submodule`, it's allowed
1648 // to access private items inside of `submodule` without a privacy
1650 super::my_implementation();
1658 For a rust program to pass the privacy checking pass, all paths must be valid
1659 accesses given the two rules above. This includes all use statements,
1660 expressions, types, etc.
1662 ### Re-exporting and Visibility
1664 Rust allows publicly re-exporting items through a `pub use` directive. Because
1665 this is a public directive, this allows the item to be used in the current
1666 module through the rules above. It essentially allows public access into the
1667 re-exported item. For example, this program is valid:
1670 pub use api = self::implementation;
1672 mod implementation {
1679 This means that any external crate referencing `implementation::f` would receive
1680 a privacy violation, while the path `api::f` would be allowed.
1682 When re-exporting a private item, it can be thought of as allowing the "privacy
1683 chain" being short-circuited through the reexport instead of passing through the
1684 namespace hierarchy as it normally would.
1686 ### Glob imports and Visibility
1688 Currently glob imports are considered an "experimental" language feature. For
1689 sanity purpose along with helping the implementation, glob imports will only
1690 import public items from their destination, not private items.
1692 > **Note:** This is subject to change, glob exports may be removed entirely or
1693 > they could possibly import private items for a privacy error to later be
1694 > issued if the item is used.
1699 attribute : '#' '[' attr_list ']' ;
1700 attr_list : attr [ ',' attr_list ]*
1701 attr : ident [ '=' literal
1702 | '(' attr_list ')' ] ? ;
1705 Static entities in Rust -- crates, modules and items -- may have _attributes_
1706 applied to them. ^[Attributes in Rust are modeled on Attributes in ECMA-335,
1708 An attribute is a general, free-form metadatum that is interpreted according to name, convention, and language and compiler version.
1709 Attributes may appear as any of
1711 * A single identifier, the attribute name
1712 * An identifier followed by the equals sign '=' and a literal, providing a key/value pair
1713 * An identifier followed by a parenthesized list of sub-attribute arguments
1715 Attributes terminated by a semi-colon apply to the entity that the attribute is declared
1716 within. Attributes that are not terminated by a semi-colon apply to the next entity.
1718 An example of attributes:
1721 // General metadata applied to the enclosing module or crate.
1724 // A function marked as a unit test
1730 // A conditionally-compiled module
1731 #[cfg(target_os="linux")]
1736 // A lint attribute used to suppress a warning/error
1737 #[allow(non_camel_case_types)]
1738 pub type int8_t = i8;
1741 > **Note:** In future versions of Rust, user-provided extensions to the compiler will be able to interpret attributes.
1742 > When this facility is provided, the compiler will distinguish between language-reserved and user-available attributes.
1744 At present, only the Rust compiler interprets attributes, so all attribute
1745 names are effectively reserved. Some significant attributes include:
1747 * The `doc` attribute, for documenting code in-place.
1748 * The `cfg` attribute, for conditional-compilation by build-configuration.
1749 * The `lang` attribute, for custom definitions of traits and functions that are known to the Rust compiler (see [Language items](#language-items)).
1750 * The `link` attribute, for describing linkage metadata for a crate.
1751 * The `test` attribute, for marking functions as unit tests.
1752 * The `allow`, `warn`, `forbid`, and `deny` attributes, for
1753 controlling lint checks (see [Lint check attributes](#lint-check-attributes)).
1754 * The `deriving` attribute, for automatically generating
1755 implementations of certain traits.
1756 * The `static_assert` attribute, for asserting that a static bool is true at compiletime
1757 * The `thread_local` attribute, for defining a `static mut` as a thread-local. Note that this is
1758 only a low-level building block, and is not local to a *task*, nor does it provide safety.
1760 Other attributes may be added or removed during development of the language.
1762 ### Lint check attributes
1764 A lint check names a potentially undesirable coding pattern, such as
1765 unreachable code or omitted documentation, for the static entity to
1766 which the attribute applies.
1768 For any lint check `C`:
1770 * `warn(C)` warns about violations of `C` but continues compilation,
1771 * `deny(C)` signals an error after encountering a violation of `C`,
1772 * `allow(C)` overrides the check for `C` so that violations will go
1774 * `forbid(C)` is the same as `deny(C)`, but also forbids uses of
1775 `allow(C)` within the entity.
1777 The lint checks supported by the compiler can be found via `rustc -W help`,
1778 along with their default settings.
1782 // Missing documentation is ignored here
1783 #[allow(missing_doc)]
1784 pub fn undocumented_one() -> int { 1 }
1786 // Missing documentation signals a warning here
1787 #[warn(missing_doc)]
1788 pub fn undocumented_too() -> int { 2 }
1790 // Missing documentation signals an error here
1791 #[deny(missing_doc)]
1792 pub fn undocumented_end() -> int { 3 }
1796 This example shows how one can use `allow` and `warn` to toggle
1797 a particular check on and off.
1800 #[warn(missing_doc)]
1802 #[allow(missing_doc)]
1804 // Missing documentation is ignored here
1805 pub fn undocumented_one() -> int { 1 }
1807 // Missing documentation signals a warning here,
1808 // despite the allow above.
1809 #[warn(missing_doc)]
1810 pub fn undocumented_two() -> int { 2 }
1813 // Missing documentation signals a warning here
1814 pub fn undocumented_too() -> int { 3 }
1818 This example shows how one can use `forbid` to disallow uses
1819 of `allow` for that lint check.
1822 #[forbid(missing_doc)]
1824 // Attempting to toggle warning signals an error here
1825 #[allow(missing_doc)]
1827 pub fn undocumented_too() -> int { 2 }
1833 Some primitive Rust operations are defined in Rust code,
1834 rather than being implemented directly in C or assembly language.
1835 The definitions of these operations have to be easy for the compiler to find.
1836 The `lang` attribute makes it possible to declare these operations.
1837 For example, the `str` module in the Rust standard library defines the string equality function:
1841 pub fn eq_slice(a: &str, b: &str) -> bool {
1846 The name `str_eq` has a special meaning to the Rust compiler,
1847 and the presence of this definition means that it will use this definition
1848 when generating calls to the string equality function.
1850 A complete list of the built-in language items follows:
1855 : Cannot be mutated.
1857 : Are uniquely owned.
1859 : Contain borrowed pointers.
1863 : Elements can be added (for example, integers and floats).
1865 : Elements can be subtracted.
1867 : Elements can be multiplied.
1869 : Elements have a division operation.
1871 : Elements have a remainder operation.
1873 : Elements can be negated arithmetically.
1875 : Elements can be negated logically.
1877 : Elements have an exclusive-or operation.
1879 : Elements have a bitwise `and` operation.
1881 : Elements have a bitwise `or` operation.
1883 : Elements have a left shift operation.
1885 : Elements have a right shift operation.
1887 : Elements can be indexed.
1889 : Elements can be compared for equality.
1891 : Elements have a partial ordering.
1896 : Compare two strings for equality.
1898 : Compare two owned strings for equality.
1900 : Destroy a box before freeing it.
1902 : Generically print a string representation of any type.
1904 : Abort the program with an error.
1906 : Abort the program with a bounds check error.
1908 : Allocate memory on the exchange heap.
1910 : Free memory that was allocated on the exchange heap.
1912 : Allocate memory on the managed heap.
1914 : Free memory that was allocated on the managed heap.
1916 : Create an immutable borrowed pointer to a mutable value.
1918 : Release a borrowed pointer created with `return_to_mut`
1919 `check_not_borrowed`
1920 : Fail if a value has existing borrowed pointers to it.
1922 : Return a new unique string
1923 containing a copy of the contents of a unique string.
1925 > **Note:** This list is likely to become out of date. We should auto-generate it
1926 > from `librustc/middle/lang_items.rs`.
1930 The `deriving` attribute allows certain traits to be automatically
1931 implemented for data structures. For example, the following will
1932 create an `impl` for the `Eq` and `Clone` traits for `Foo`, the type
1933 parameter `T` will be given the `Eq` or `Clone` constraints for the
1937 #[deriving(Eq, Clone)]
1944 The generated `impl` for `Eq` is equivalent to
1947 # struct Foo<T> { a: int, b: T }
1948 impl<T: Eq> Eq for Foo<T> {
1949 fn eq(&self, other: &Foo<T>) -> bool {
1950 self.a == other.a && self.b == other.b
1953 fn ne(&self, other: &Foo<T>) -> bool {
1954 self.a != other.a || self.b != other.b
1959 Supported traits for `deriving` are:
1961 * Comparison traits: `Eq`, `TotalEq`, `Ord`, `TotalOrd`.
1962 * Serialization: `Encodable`, `Decodable`. These require `extra`.
1963 * `Clone` and `DeepClone`, to perform (deep) copies.
1964 * `IterBytes`, to iterate over the bytes in a data type.
1965 * `Rand`, to create a random instance of a data type.
1966 * `Default`, to create an empty instance of a data type.
1967 * `Zero`, to create an zero instance of a numeric data type.
1968 * `ToStr`, to convert to a string. For a type with this instance,
1969 `obj.to_str()` has similar output as `fmt!("%?", obj)`, but it differs in that
1970 each constituent field of the type must also implement `ToStr` and will have
1971 `field.to_str()` invoked to build up the result.
1972 * `FromPrimitive`, to create an instance from a numeric primitve.
1975 One can indicate the stability of an API using the following attributes:
1977 * `deprecated`: This item should no longer be used, e.g. it has been
1978 replaced. No guarantee of backwards-compatibility.
1979 * `experimental`: This item was only recently introduced or is
1980 otherwise in a state of flux. It may change significantly, or even
1981 be removed. No guarantee of backwards-compatibility.
1982 * `unstable`: This item is still under development, but requires more
1983 testing to be considered stable. No guarantee of backwards-compatibility.
1984 * `stable`: This item is considered stable, and will not change
1985 significantly. Guarantee of backwards-compatibility.
1986 * `frozen`: This item is very stable, and is unlikely to
1987 change. Guarantee of backwards-compatibility.
1988 * `locked`: This item will never change unless a serious bug is
1989 found. Guarantee of backwards-compatibility.
1991 These levels are directly inspired by
1992 [Node.js' "stability index"](http://nodejs.org/api/documentation.html).
1994 There are lints for disallowing items marked with certain levels:
1995 `deprecated`, `experimental` and `unstable`; the first two will warn
1996 by default. Items with not marked with a stability are considered to
1997 be unstable for the purposes of the lint. One can give an optional
1998 string that will be displayed when the lint flags the use of an item.
2003 #[deprecated="replaced by `best`"]
2005 // delete everything
2009 // delete fewer things
2018 bad(); // "warning: use of deprecated item: replaced by `best`"
2020 better(); // "warning: use of unmarked item"
2022 best(); // no warning
2026 > **Note:** Currently these are only checked when applied to
2027 > individual functions, structs, methods and enum variants, *not* to
2028 > entire modules, traits, impls or enums themselves.
2030 ### Compiler Features
2032 Certain aspects of Rust may be implemented in the compiler, but they're not
2033 necessarily ready for every-day use. These features are often of "prototype
2034 quality" or "almost production ready", but may not be stable enough to be
2035 considered a full-fleged language feature.
2037 For this reason, rust recognizes a special crate-level attribute of the form:
2040 #[feature(feature1, feature2, feature3)]
2043 This directive informs the compiler that the feature list: `feature1`,
2044 `feature2`, and `feature3` should all be enabled. This is only recognized at a
2045 crate-level, not at a module-level. Without this directive, all features are
2046 considered off, and using the features will result in a compiler error.
2048 The currently implemented features of the compiler are:
2050 * `macro_rules` - The definition of new macros. This does not encompass
2051 macro-invocation, that is always enabled by default, this only
2052 covers the definition of new macros. There are currently
2053 various problems with invoking macros, how they interact with
2054 their environment, and possibly how they are used outside of
2055 location in which they are defined. Macro definitions are
2056 likely to change slightly in the future, so they are currently
2057 hidden behind this feature.
2059 * `globs` - Importing everything in a module through `*`. This is currently a
2060 large source of bugs in name resolution for Rust, and it's not clear
2061 whether this will continue as a feature or not. For these reasons,
2062 the glob import statement has been hidden behind this feature flag.
2064 * `struct_variant` - Structural enum variants (those with named fields). It is
2065 currently unknown whether this style of enum variant is as
2066 fully supported as the tuple-forms, and it's not certain
2067 that this style of variant should remain in the language.
2068 For now this style of variant is hidden behind a feature
2071 * `once_fns` - Onceness guarantees a closure is only executed once. Defining a
2072 closure as `once` is unlikely to be supported going forward. So
2073 they are hidden behind this feature until they are to be removed.
2075 If a feature is promoted to a language feature, then all existing programs will
2076 start to receive compilation warnings about #[feature] directives which enabled
2077 the new feature (because the directive is no longer necessary). However, if
2078 a feature is decided to be removed from the language, errors will be issued (if
2079 there isn't a parser error first). The directive in this case is no longer
2080 necessary, and it's likely that existing code will break if the feature isn't
2083 If a unknown feature is found in a directive, it results in a compiler error. An
2084 unknown feature is one which has never been recognized by the compiler.
2086 # Statements and expressions
2088 Rust is _primarily_ an expression language. This means that most forms of
2089 value-producing or effect-causing evaluation are directed by the uniform
2090 syntax category of _expressions_. Each kind of expression can typically _nest_
2091 within each other kind of expression, and rules for evaluation of expressions
2092 involve specifying both the value produced by the expression and the order in
2093 which its sub-expressions are themselves evaluated.
2095 In contrast, statements in Rust serve _mostly_ to contain and explicitly
2096 sequence expression evaluation.
2100 A _statement_ is a component of a block, which is in turn a component of an
2101 outer [expression](#expressions) or [function](#functions).
2103 Rust has two kinds of statement:
2104 [declaration statements](#declaration-statements) and
2105 [expression statements](#expression-statements).
2107 ### Declaration statements
2109 A _declaration statement_ is one that introduces one or more *names* into the enclosing statement block.
2110 The declared names may denote new slots or new items.
2112 #### Item declarations
2114 An _item declaration statement_ has a syntactic form identical to an
2115 [item](#items) declaration within a module. Declaring an item -- a function,
2116 enumeration, structure, type, static, trait, implementation or module -- locally
2117 within a statement block is simply a way of restricting its scope to a narrow
2118 region containing all of its uses; it is otherwise identical in meaning to
2119 declaring the item outside the statement block.
2121 Note: there is no implicit capture of the function's dynamic environment when
2122 declaring a function-local item.
2124 #### Slot declarations
2127 let_decl : "let" pat [':' type ] ? [ init ] ? ';' ;
2128 init : [ '=' ] expr ;
2131 A _slot declaration_ introduces a new set of slots, given by a pattern.
2132 The pattern may be followed by a type annotation, and/or an initializer expression.
2133 When no type annotation is given, the compiler will infer the type,
2134 or signal an error if insufficient type information is available for definite inference.
2135 Any slots introduced by a slot declaration are visible from the point of declaration until the end of the enclosing block scope.
2137 ### Expression statements
2139 An _expression statement_ is one that evaluates an [expression](#expressions)
2140 and ignores its result.
2141 The type of an expression statement `e;` is always `()`, regardless of the type of `e`.
2142 As a rule, an expression statement's purpose is to trigger the effects of evaluating its expression.
2146 An expression may have two roles: it always produces a *value*, and it may have *effects*
2147 (otherwise known as "side effects").
2148 An expression *evaluates to* a value, and has effects during *evaluation*.
2149 Many expressions contain sub-expressions (operands).
2150 The meaning of each kind of expression dictates several things:
2151 * Whether or not to evaluate the sub-expressions when evaluating the expression
2152 * The order in which to evaluate the sub-expressions
2153 * How to combine the sub-expressions' values to obtain the value of the expression.
2155 In this way, the structure of expressions dictates the structure of execution.
2156 Blocks are just another kind of expression,
2157 so blocks, statements, expressions, and blocks again can recursively nest inside each other
2158 to an arbitrary depth.
2160 #### Lvalues, rvalues and temporaries
2162 Expressions are divided into two main categories: _lvalues_ and _rvalues_.
2163 Likewise within each expression, sub-expressions may occur in _lvalue context_ or _rvalue context_.
2164 The evaluation of an expression depends both on its own category and the context it occurs within.
2166 [Path](#path-expressions), [field](#field-expressions) and [index](#index-expressions) expressions are lvalues.
2167 All other expressions are rvalues.
2169 The left operand of an [assignment](#assignment-expressions),
2170 [binary move](#binary-move-expressions) or
2171 [compound-assignment](#compound-assignment-expressions) expression is an lvalue context,
2172 as is the single operand of a unary [borrow](#unary-operator-expressions),
2173 or [move](#unary-move-expressions) expression,
2174 and _both_ operands of a [swap](#swap-expressions) expression.
2175 All other expression contexts are rvalue contexts.
2177 When an lvalue is evaluated in an _lvalue context_, it denotes a memory location;
2178 when evaluated in an _rvalue context_, it denotes the value held _in_ that memory location.
2180 When an rvalue is used in lvalue context, a temporary un-named lvalue is created and used instead.
2181 A temporary's lifetime equals the largest lifetime of any borrowed pointer that points to it.
2183 #### Moved and copied types
2185 When a [local variable](#memory-slots) is used
2186 as an [rvalue](#lvalues-rvalues-and-temporaries)
2187 the variable will either be [moved](#move-expressions) or copied,
2188 depending on its type.
2189 For types that contain [owning pointers](#owning-pointers)
2190 or values that implement the special trait `Drop`,
2191 the variable is moved.
2192 All other types are copied.
2194 ### Literal expressions
2196 A _literal expression_ consists of one of the [literal](#literals)
2197 forms described earlier. It directly describes a number, character,
2198 string, boolean value, or the unit value.
2202 "hello"; // string type
2203 '5'; // character type
2207 ### Path expressions
2209 A [path](#paths) used as an expression context denotes either a local variable or an item.
2210 Path expressions are [lvalues](#lvalues-rvalues-and-temporaries).
2212 ### Tuple expressions
2214 Tuples are written by enclosing one or more comma-separated
2215 expressions in parentheses. They are used to create [tuple-typed](#tuple-types)
2224 ### Structure expressions
2227 struct_expr : expr_path '{' ident ':' expr
2228 [ ',' ident ':' expr ] *
2231 [ ',' expr ] * ')' |
2235 There are several forms of structure expressions.
2236 A _structure expression_ consists of the [path](#paths) of a [structure item](#structures),
2237 followed by a brace-enclosed list of one or more comma-separated name-value pairs,
2238 providing the field values of a new instance of the structure.
2239 A field name can be any identifier, and is separated from its value expression by a colon.
2240 The location denoted by a structure field is mutable if and only if the enclosing structure is mutable.
2242 A _tuple structure expression_ consists of the [path](#paths) of a [structure item](#structures),
2243 followed by a parenthesized list of one or more comma-separated expressions
2244 (in other words, the path of a structure item followed by a tuple expression).
2245 The structure item must be a tuple structure item.
2247 A _unit-like structure expression_ consists only of the [path](#paths) of a [structure item](#structures).
2249 The following are examples of structure expressions:
2252 # struct Point { x: f64, y: f64 }
2253 # struct TuplePoint(f64, f64);
2254 # mod game { pub struct User<'self> { name: &'self str, age: uint, score: uint } }
2255 # struct Cookie; fn some_fn<T>(t: T) {}
2256 Point {x: 10.0, y: 20.0};
2257 TuplePoint(10.0, 20.0);
2258 let u = game::User {name: "Joe", age: 35, score: 100_000};
2259 some_fn::<Cookie>(Cookie);
2262 A structure expression forms a new value of the named structure type.
2263 Note that for a given *unit-like* structure type, this will always be the same value.
2265 A structure expression can terminate with the syntax `..` followed by an expression to denote a functional update.
2266 The expression following `..` (the base) must have the same structure type as the new structure type being formed.
2267 The entire expression denotes the result of allocating a new structure
2268 (with the same type as the base expression)
2269 with the given values for the fields that were explicitly specified
2270 and the values in the base record for all other fields.
2273 # struct Point3d { x: int, y: int, z: int }
2274 let base = Point3d {x: 1, y: 2, z: 3};
2275 Point3d {y: 0, z: 10, .. base};
2278 ### Record expressions
2281 rec_expr : '{' ident ':' expr
2282 [ ',' ident ':' expr ] *
2286 ### Method-call expressions
2289 method_call_expr : expr '.' ident paren_expr_list ;
2292 A _method call_ consists of an expression followed by a single dot, an identifier, and a parenthesized expression-list.
2293 Method calls are resolved to methods on specific traits,
2294 either statically dispatching to a method if the exact `self`-type of the left-hand-side is known,
2295 or dynamically dispatching if the left-hand-side expression is an indirect [object type](#object-types).
2297 ### Field expressions
2300 field_expr : expr '.' ident
2303 A _field expression_ consists of an expression followed by a single dot and an identifier,
2304 when not immediately followed by a parenthesized expression-list (the latter is a [method call expression](#method-call-expressions)).
2305 A field expression denotes a field of a [structure](#structure-types).
2312 A field access on a record is an [lvalue](#lvalues-rvalues-and-temporaries) referring to the value of that field.
2313 When the field is mutable, it can be [assigned](#assignment-expressions) to.
2315 When the type of the expression to the left of the dot is a pointer to a record or structure,
2316 it is automatically dereferenced to make the field access possible.
2318 ### Vector expressions
2321 vec_expr : '[' "mut" ? vec_elems? ']'
2323 vec_elems : [expr [',' expr]*] | [expr ',' ".." expr]
2326 A [_vector_](#vector-types) _expression_ is written by enclosing zero or
2327 more comma-separated expressions of uniform type in square brackets.
2329 In the `[expr ',' ".." expr]` form, the expression after the `".."`
2330 must be a constant expression that can be evaluated at compile time, such
2331 as a [literal](#literals) or a [static item](#static-items).
2335 ["a", "b", "c", "d"];
2336 [0, ..128]; // vector with 128 zeros
2337 [0u8, 0u8, 0u8, 0u8];
2340 ### Index expressions
2343 idx_expr : expr '[' expr ']'
2346 [Vector](#vector-types)-typed expressions can be indexed by writing a
2347 square-bracket-enclosed expression (the index) after them. When the
2348 vector is mutable, the resulting [lvalue](#lvalues-rvalues-and-temporaries) can be assigned to.
2350 Indices are zero-based, and may be of any integral type. Vector access
2351 is bounds-checked at run-time. When the check fails, it will put the
2352 task in a _failing state_.
2359 (["a", "b"])[10]; // fails
2364 ### Unary operator expressions
2366 Rust defines six symbolic unary operators.
2367 They are all written as prefix operators,
2368 before the expression they apply to.
2371 : Negation. May only be applied to numeric types.
2373 : Dereference. When applied to a [pointer](#pointer-types) it denotes the pointed-to location.
2374 For pointers to mutable locations, the resulting [lvalue](#lvalues-rvalues-and-temporaries) can be assigned to.
2375 For [enums](#enumerated-types) that have only a single variant, containing a single parameter,
2376 the dereference operator accesses this parameter.
2378 : Logical negation. On the boolean type, this flips between `true` and
2379 `false`. On integer types, this inverts the individual bits in the
2380 two's complement representation of the value.
2382 : [Boxing](#pointer-types) operators. Allocate a box to hold the value they are applied to,
2383 and store the value in it. `@` creates a managed box, whereas `~` creates an owned box.
2385 : Borrow operator. Returns a borrowed pointer, pointing to its operand.
2386 The operand of a borrowed pointer is statically proven to outlive the resulting pointer.
2387 If the borrow-checker cannot prove this, it is a compilation error.
2389 ### Binary operator expressions
2392 binop_expr : expr binop expr ;
2395 Binary operators expressions are given in terms of
2396 [operator precedence](#operator-precedence).
2398 #### Arithmetic operators
2400 Binary arithmetic expressions are syntactic sugar for calls to built-in traits,
2401 defined in the `std::ops` module of the `std` library.
2402 This means that arithmetic operators can be overridden for user-defined types.
2403 The default meaning of the operators on standard types is given here.
2406 : Addition and vector/string concatenation.
2407 Calls the `add` method on the `std::ops::Add` trait.
2410 Calls the `sub` method on the `std::ops::Sub` trait.
2413 Calls the `mul` method on the `std::ops::Mul` trait.
2416 Calls the `div` method on the `std::ops::Div` trait.
2419 Calls the `rem` method on the `std::ops::Rem` trait.
2421 #### Bitwise operators
2423 Like the [arithmetic operators](#arithmetic-operators), bitwise operators
2424 are syntactic sugar for calls to methods of built-in traits.
2425 This means that bitwise operators can be overridden for user-defined types.
2426 The default meaning of the operators on standard types is given here.
2430 Calls the `bitand` method of the `std::ops::BitAnd` trait.
2433 Calls the `bitor` method of the `std::ops::BitOr` trait.
2436 Calls the `bitxor` method of the `std::ops::BitXor` trait.
2438 : Logical left shift.
2439 Calls the `shl` method of the `std::ops::Shl` trait.
2441 : Logical right shift.
2442 Calls the `shr` method of the `std::ops::Shr` trait.
2444 #### Lazy boolean operators
2446 The operators `||` and `&&` may be applied to operands of boolean type.
2447 The `||` operator denotes logical 'or', and the `&&` operator denotes logical 'and'.
2448 They differ from `|` and `&` in that the right-hand operand is only evaluated
2449 when the left-hand operand does not already determine the result of the expression.
2450 That is, `||` only evaluates its right-hand operand
2451 when the left-hand operand evaluates to `false`, and `&&` only when it evaluates to `true`.
2453 #### Comparison operators
2455 Comparison operators are, like the [arithmetic operators](#arithmetic-operators),
2456 and [bitwise operators](#bitwise-operators),
2457 syntactic sugar for calls to built-in traits.
2458 This means that comparison operators can be overridden for user-defined types.
2459 The default meaning of the operators on standard types is given here.
2463 Calls the `eq` method on the `std::cmp::Eq` trait.
2466 Calls the `ne` method on the `std::cmp::Eq` trait.
2469 Calls the `lt` method on the `std::cmp::Ord` trait.
2472 Calls the `gt` method on the `std::cmp::Ord` trait.
2474 : Less than or equal.
2475 Calls the `le` method on the `std::cmp::Ord` trait.
2477 : Greater than or equal.
2478 Calls the `ge` method on the `std::cmp::Ord` trait.
2480 #### Type cast expressions
2482 A type cast expression is denoted with the binary operator `as`.
2484 Executing an `as` expression casts the value on the left-hand side to the type
2485 on the right-hand side.
2487 A numeric value can be cast to any numeric type.
2488 A raw pointer value can be cast to or from any integral type or raw pointer type.
2489 Any other cast is unsupported and will fail to compile.
2491 An example of an `as` expression:
2494 # fn sum(v: &[f64]) -> f64 { 0.0 }
2495 # fn len(v: &[f64]) -> int { 0 }
2497 fn avg(v: &[f64]) -> f64 {
2498 let sum: f64 = sum(v);
2499 let sz: f64 = len(v) as f64;
2504 #### Assignment expressions
2506 An _assignment expression_ consists of an [lvalue](#lvalues-rvalues-and-temporaries) expression followed by an
2507 equals sign (`=`) and an [rvalue](#lvalues-rvalues-and-temporaries) expression.
2509 Evaluating an assignment expression [either copies or moves](#moved-and-copied-types) its right-hand operand to its left-hand operand.
2518 #### Compound assignment expressions
2520 The `+`, `-`, `*`, `/`, `%`, `&`, `|`, `^`, `<<`, and `>>`
2521 operators may be composed with the `=` operator. The expression `lval
2522 OP= val` is equivalent to `lval = lval OP val`. For example, `x = x +
2523 1` may be written as `x += 1`.
2525 Any such expression always has the [`unit`](#primitive-types) type.
2527 #### Operator precedence
2529 The precedence of Rust binary operators is ordered as follows, going
2530 from strong to weak:
2547 Operators at the same precedence level are evaluated left-to-right. [Unary operators](#unary-operator-expressions)
2548 have the same precedence level and it is stronger than any of the binary operators'.
2550 ### Grouped expressions
2552 An expression enclosed in parentheses evaluates to the result of the enclosed
2553 expression. Parentheses can be used to explicitly specify evaluation order
2554 within an expression.
2557 paren_expr : '(' expr ')' ;
2560 An example of a parenthesized expression:
2563 let x = (2 + 3) * 4;
2567 ### Call expressions
2570 expr_list : [ expr [ ',' expr ]* ] ? ;
2571 paren_expr_list : '(' expr_list ')' ;
2572 call_expr : expr paren_expr_list ;
2575 A _call expression_ invokes a function, providing zero or more input slots and
2576 an optional reference slot to serve as the function's output, bound to the
2577 `lval` on the right hand side of the call. If the function eventually returns,
2578 then the expression completes.
2580 Some examples of call expressions:
2583 # use std::from_str::FromStr;
2584 # fn add(x: int, y: int) -> int { 0 }
2586 let x: int = add(1, 2);
2587 let pi: Option<f32> = FromStr::from_str("3.14");
2590 ### Lambda expressions
2593 ident_list : [ ident [ ',' ident ]* ] ? ;
2594 lambda_expr : '|' ident_list '|' expr ;
2597 A _lambda expression_ (sometimes called an "anonymous function expression") defines a function and denotes it as a value,
2598 in a single expression.
2599 A lambda expression is a pipe-symbol-delimited (`|`) list of identifiers followed by an expression.
2601 A lambda expression denotes a function that maps a list of parameters (`ident_list`)
2602 onto the expression that follows the `ident_list`.
2603 The identifiers in the `ident_list` are the parameters to the function.
2604 These parameters' types need not be specified, as the compiler infers them from context.
2606 Lambda expressions are most useful when passing functions as arguments to other functions,
2607 as an abbreviation for defining and capturing a separate function.
2609 Significantly, lambda expressions _capture their environment_,
2610 which regular [function definitions](#functions) do not.
2611 The exact type of capture depends on the [function type](#function-types) inferred for the lambda expression.
2612 In the simplest and least-expensive form (analogous to a ```|| { }``` expression),
2613 the lambda expression captures its environment by reference,
2614 effectively borrowing pointers to all outer variables mentioned inside the function.
2615 Alternately, the compiler may infer that a lambda expression should copy or move values (depending on their type.)
2616 from the environment into the lambda expression's captured environment.
2618 In this example, we define a function `ten_times` that takes a higher-order function argument,
2619 and call it with a lambda expression as an argument.
2622 fn ten_times(f: |int|) {
2630 ten_times(|j| println!("hello, {}", j));
2636 while_expr : "while" expr '{' block '}' ;
2639 A `while` loop begins by evaluating the boolean loop conditional expression.
2640 If the loop conditional expression evaluates to `true`, the loop body block
2641 executes and control returns to the loop conditional expression. If the loop
2642 conditional expression evaluates to `false`, the `while` expression completes.
2657 The keyword `loop` in Rust appears both in _loop expressions_ and in _continue expressions_.
2658 A loop expression denotes an infinite loop;
2659 see [Continue expressions](#continue-expressions) for continue expressions.
2662 loop_expr : [ lifetime ':' ] "loop" '{' block '}';
2665 A `loop` expression may optionally have a _label_.
2666 If a label is present,
2667 then labeled `break` and `loop` expressions nested within this loop may exit out of this loop or return control to its head.
2668 See [Break expressions](#break-expressions).
2670 ### Break expressions
2673 break_expr : "break" [ lifetime ];
2676 A `break` expression has an optional `label`.
2677 If the label is absent, then executing a `break` expression immediately terminates the innermost loop enclosing it.
2678 It is only permitted in the body of a loop.
2679 If the label is present, then `break foo` terminates the loop with label `foo`,
2680 which need not be the innermost label enclosing the `break` expression,
2681 but must enclose it.
2683 ### Continue expressions
2686 continue_expr : "loop" [ lifetime ];
2689 A continue expression, written `loop`, also has an optional `label`.
2690 If the label is absent,
2691 then executing a `loop` expression immediately terminates the current iteration of the innermost loop enclosing it,
2692 returning control to the loop *head*.
2693 In the case of a `while` loop,
2694 the head is the conditional expression controlling the loop.
2695 In the case of a `for` loop, the head is the call-expression controlling the loop.
2696 If the label is present, then `loop foo` returns control to the head of the loop with label `foo`,
2697 which need not be the innermost label enclosing the `break` expression,
2698 but must enclose it.
2700 A `loop` expression is only permitted in the body of a loop.
2705 do_expr : "do" expr [ '|' ident_list '|' ] ? '{' block '}' ;
2708 A _do expression_ provides a more-familiar block-syntax for a [lambda expression](#lambda-expressions),
2709 including a special translation of [return expressions](#return-expressions) inside the supplied block.
2711 Any occurrence of a [return expression](#return-expressions)
2712 inside this `block` expression is rewritten
2713 as a reference to an (anonymous) flag set in the caller's environment,
2714 which is checked on return from the `expr` and, if set,
2715 causes a corresponding return from the caller.
2716 In this way, the meaning of `return` statements in language built-in control blocks is preserved,
2717 if they are rewritten using lambda functions and `do` expressions as abstractions.
2719 The optional `ident_list` and `block` provided in a `do` expression are parsed as though they constitute a lambda expression;
2720 if the `ident_list` is missing, an empty `ident_list` is implied.
2722 The lambda expression is then provided as a _trailing argument_
2723 to the outermost [call](#call-expressions) or [method call](#method-call-expressions) expression
2724 in the `expr` following `do`.
2725 If the `expr` is a [path expression](#path-expressions), it is parsed as though it is a call expression.
2726 If the `expr` is a [field expression](#field-expressions), it is parsed as though it is a method call expression.
2728 In this example, both calls to `f` are equivalent:
2731 # fn f(f: |int|) { }
2741 In this example, both calls to the (binary) function `k` are equivalent:
2744 # fn k(x:int, f: |int|) { }
2757 for_expr : "for" pat "in" expr '{' block '}' ;
2760 A `for` expression is a syntactic construct for looping over elements
2761 provided by an implementation of `std::iter::Iterator`.
2763 An example of a for loop over the contents of a vector:
2767 # fn bar(f: foo) { }
2772 let v: &[foo] = &[a, b, c];
2779 An example of a for loop over a series of integers:
2782 # fn bar(b:uint) { }
2783 for i in range(0u, 256) {
2791 if_expr : "if" expr '{' block '}'
2794 else_tail : "else" [ if_expr
2798 An `if` expression is a conditional branch in program control. The form of
2799 an `if` expression is a condition expression, followed by a consequent
2800 block, any number of `else if` conditions and blocks, and an optional
2801 trailing `else` block. The condition expressions must have type
2802 `bool`. If a condition expression evaluates to `true`, the
2803 consequent block is executed and any subsequent `else if` or `else`
2804 block is skipped. If a condition expression evaluates to `false`, the
2805 consequent block is skipped and any subsequent `else if` condition is
2806 evaluated. If all `if` and `else if` conditions evaluate to `false`
2807 then any `else` block is executed.
2809 ### Match expressions
2812 match_expr : "match" expr '{' match_arm [ '|' match_arm ] * '}' ;
2814 match_arm : match_pat '=>' [ expr "," | '{' block '}' ] ;
2816 match_pat : pat [ ".." pat ] ? [ "if" expr ] ;
2819 A `match` expression branches on a *pattern*. The exact form of matching that
2820 occurs depends on the pattern. Patterns consist of some combination of
2821 literals, destructured enum constructors, structures, records and tuples, variable binding
2822 specifications, wildcards (`*`), and placeholders (`_`). A `match` expression has a *head
2823 expression*, which is the value to compare to the patterns. The type of the
2824 patterns must equal the type of the head expression.
2826 In a pattern whose head expression has an `enum` type, a placeholder (`_`) stands for a
2827 *single* data field, whereas a wildcard `*` stands for *all* the fields of a particular
2828 variant. For example:
2831 enum List<X> { Nil, Cons(X, @List<X>) }
2833 let x: List<int> = Cons(10, @Cons(11, @Nil));
2836 Cons(_, @Nil) => fail!("singleton list"),
2838 Nil => fail!("empty list")
2842 The first pattern matches lists constructed by applying `Cons` to any head value, and a
2843 tail value of `@Nil`. The second pattern matches _any_ list constructed with `Cons`,
2844 ignoring the values of its arguments. The difference between `_` and `*` is that the pattern `C(_)` is only type-correct if
2845 `C` has exactly one argument, while the pattern `C(*)` is type-correct for any enum variant `C`, regardless of how many arguments `C` has.
2847 To execute an `match` expression, first the head expression is evaluated, then
2848 its value is sequentially compared to the patterns in the arms until a match
2849 is found. The first arm with a matching pattern is chosen as the branch target
2850 of the `match`, any variables bound by the pattern are assigned to local
2851 variables in the arm's block, and control enters the block.
2853 An example of an `match` expression:
2856 # fn process_pair(a: int, b: int) { }
2857 # fn process_ten() { }
2859 enum List<X> { Nil, Cons(X, @List<X>) }
2861 let x: List<int> = Cons(10, @Cons(11, @Nil));
2864 Cons(a, @Cons(b, _)) => {
2879 Patterns that bind variables
2880 default to binding to a copy or move of the matched value
2881 (depending on the matched value's type).
2882 This can be changed to bind to a borrowed pointer by
2883 using the ```ref``` keyword,
2884 or to a mutable borrowed pointer using ```ref mut```.
2886 A pattern that's just an identifier,
2887 like `Nil` in the previous answer,
2888 could either refer to an enum variant that's in scope,
2889 or bind a new variable.
2890 The compiler resolves this ambiguity by forbidding variable bindings that occur in ```match``` patterns from shadowing names of variants that are in scope.
2891 For example, wherever ```List``` is in scope,
2892 a ```match``` pattern would not be able to bind ```Nil``` as a new name.
2893 The compiler interprets a variable pattern `x` as a binding _only_ if there is no variant named `x` in scope.
2894 A convention you can use to avoid conflicts is simply to name variants with upper-case letters,
2895 and local variables with lower-case letters.
2897 Multiple match patterns may be joined with the `|` operator.
2898 A range of values may be specified with `..`.
2904 let message = match x {
2905 0 | 1 => "not many",
2911 Range patterns only work on scalar types
2912 (like integers and characters; not like vectors and structs, which have sub-components).
2913 A range pattern may not be a sub-range of another range pattern inside the same `match`.
2915 Finally, match patterns can accept *pattern guards* to further refine the
2916 criteria for matching a case. Pattern guards appear after the pattern and
2917 consist of a bool-typed expression following the `if` keyword. A pattern
2918 guard may refer to the variables bound within the pattern they follow.
2921 # let maybe_digit = Some(0);
2922 # fn process_digit(i: int) { }
2923 # fn process_other(i: int) { }
2925 let message = match maybe_digit {
2926 Some(x) if x < 10 => process_digit(x),
2927 Some(x) => process_other(x),
2932 ### Return expressions
2935 return_expr : "return" expr ? ;
2938 Return expressions are denoted with the keyword `return`. Evaluating a `return`
2939 expression moves its argument into the output slot of the current
2940 function, destroys the current function activation frame, and transfers
2941 control to the caller frame.
2943 An example of a `return` expression:
2946 fn max(a: int, b: int) -> int {
2958 Every slot, item and value in a Rust program has a type. The _type_ of a *value*
2959 defines the interpretation of the memory holding it.
2961 Built-in types and type-constructors are tightly integrated into the language,
2962 in nontrivial ways that are not possible to emulate in user-defined
2963 types. User-defined types have limited capabilities.
2967 The primitive types are the following:
2969 * The "unit" type `()`, having the single "unit" value `()` (occasionally called "nil").
2970 ^[The "unit" value `()` is *not* a sentinel "null pointer" value for reference slots; the "unit" type is the implicit return type from functions otherwise lacking a return type, and can be used in other contexts (such as message-sending or type-parametric code) as a zero-size type.]
2971 * The boolean type `bool` with values `true` and `false`.
2972 * The machine types.
2973 * The machine-dependent integer and floating-point types.
2977 The machine types are the following:
2979 * The unsigned word types `u8`, `u16`, `u32` and `u64`, with values drawn from
2980 the integer intervals $[0, 2^8 - 1]$, $[0, 2^{16} - 1]$, $[0, 2^{32} - 1]$ and
2981 $[0, 2^{64} - 1]$ respectively.
2983 * The signed two's complement word types `i8`, `i16`, `i32` and `i64`, with
2984 values drawn from the integer intervals $[-(2^7), 2^7 - 1]$,
2985 $[-(2^{15}), 2^{15} - 1]$, $[-(2^{31}), 2^{31} - 1]$, $[-(2^{63}), 2^{63} - 1]$
2988 * The IEEE 754-2008 `binary32` and `binary64` floating-point types: `f32` and
2989 `f64`, respectively.
2991 #### Machine-dependent integer types
2993 The Rust type `uint`^[A Rust `uint` is analogous to a C99 `uintptr_t`.] is an
2994 unsigned integer type with target-machine-dependent size. Its size, in
2995 bits, is equal to the number of bits required to hold any memory address on
2998 The Rust type `int`^[A Rust `int` is analogous to a C99 `intptr_t`.] is a
2999 two's complement signed integer type with target-machine-dependent size. Its
3000 size, in bits, is equal to the size of the rust type `uint` on the same target
3005 The types `char` and `str` hold textual data.
3007 A value of type `char` is a Unicode character,
3008 represented as a 32-bit unsigned word holding a UCS-4 codepoint.
3010 A value of type `str` is a Unicode string,
3011 represented as a vector of 8-bit unsigned bytes holding a sequence of UTF-8 codepoints.
3012 Since `str` is of unknown size, it is not a _first class_ type,
3013 but can only be instantiated through a pointer type,
3014 such as `&str`, `@str` or `~str`.
3018 The tuple type-constructor forms a new heterogeneous product of values similar
3019 to the record type-constructor. The differences are as follows:
3021 * tuple elements cannot be mutable, unlike record fields
3022 * tuple elements are not named and can be accessed only by pattern-matching
3024 Tuple types and values are denoted by listing the types or values of their
3025 elements, respectively, in a parenthesized, comma-separated
3028 The members of a tuple are laid out in memory contiguously, like a record, in
3029 order specified by the tuple type.
3031 An example of a tuple type and its use:
3034 type Pair<'self> = (int,&'self str);
3035 let p: Pair<'static> = (10,"hello");
3037 assert!(b != "world");
3042 The vector type constructor represents a homogeneous array of values of a given type.
3043 A vector has a fixed size.
3044 (Operations like `vec.push` operate solely on owned vectors.)
3045 A vector type can be annotated with a _definite_ size,
3046 written with a trailing asterisk and integer literal, such as `[int * 10]`.
3047 Such a definite-sized vector type is a first-class type, since its size is known statically.
3048 A vector without such a size is said to be of _indefinite_ size,
3049 and is therefore not a _first-class_ type.
3050 An indefinite-size vector can only be instantiated through a pointer type,
3051 such as `&[T]`, `@[T]` or `~[T]`.
3052 The kind of a vector type depends on the kind of its element type,
3053 as with other simple structural types.
3055 Expressions producing vectors of definite size cannot be evaluated in a
3056 context expecting a vector of indefinite size; one must copy the
3057 definite-sized vector contents into a distinct vector of indefinite size.
3059 An example of a vector type and its use:
3062 let v: &[int] = &[7, 5, 3];
3067 All in-bounds elements of a vector are always initialized,
3068 and access to a vector is always bounds-checked.
3072 A `struct` *type* is a heterogeneous product of other types, called the *fields* of the type.
3073 ^[`struct` types are analogous `struct` types in C,
3074 the *record* types of the ML family,
3075 or the *structure* types of the Lisp family.]
3077 New instances of a `struct` can be constructed with a [struct expression](#struct-expressions).
3079 The memory order of fields in a `struct` is given by the item defining it.
3080 Fields may be given in any order in a corresponding struct *expression*;
3081 the resulting `struct` value will always be laid out in memory in the order specified by the corresponding *item*.
3083 The fields of a `struct` may be qualified by [visibility modifiers](#visibility-modifiers),
3084 to restrict access to implementation-private data in a structure.
3086 A _tuple struct_ type is just like a structure type, except that the fields are anonymous.
3088 A _unit-like struct_ type is like a structure type, except that it has no fields.
3089 The one value constructed by the associated [structure expression](#structure-expression) is the only value that inhabits such a type.
3091 ### Enumerated types
3093 An *enumerated type* is a nominal, heterogeneous disjoint union type,
3094 denoted by the name of an [`enum` item](#enumerations).
3095 ^[The `enum` type is analogous to a `data` constructor declaration in ML,
3096 or a *pick ADT* in Limbo.]
3098 An [`enum` item](#enumerations) declares both the type and a number of *variant constructors*,
3099 each of which is independently named and takes an optional tuple of arguments.
3101 New instances of an `enum` can be constructed by calling one of the variant constructors,
3102 in a [call expression](#call-expressions).
3104 Any `enum` value consumes as much memory as the largest variant constructor for its corresponding `enum` type.
3106 Enum types cannot be denoted *structurally* as types,
3107 but must be denoted by named reference to an [`enum` item](#enumerations).
3111 Nominal types -- [enumerations](#enumerated-types) and [structures](#structure-types) -- may be recursive.
3112 That is, each `enum` constructor or `struct` field may refer, directly or indirectly, to the enclosing `enum` or `struct` type itself.
3113 Such recursion has restrictions:
3115 * Recursive types must include a nominal type in the recursion
3116 (not mere [type definitions](#type-definitions),
3117 or other structural types such as [vectors](#vector-types) or [tuples](#tuple-types)).
3118 * A recursive `enum` item must have at least one non-recursive constructor
3119 (in order to give the recursion a basis case).
3120 * The size of a recursive type must be finite;
3121 in other words the recursive fields of the type must be [pointer types](#pointer-types).
3122 * Recursive type definitions can cross module boundaries, but not module *visibility* boundaries,
3123 or crate boundaries (in order to simplify the module system and type checker).
3125 An example of a *recursive* type and its use:
3133 let a: List<int> = Cons(7, @Cons(13, @Nil));
3138 All pointers in Rust are explicit first-class values.
3139 They can be copied, stored into data structures, and returned from functions.
3140 There are four varieties of pointer in Rust:
3142 Managed pointers (`@`)
3143 : These point to managed heap allocations (or "boxes") in the task-local, managed heap.
3144 Managed pointers are written `@content`,
3145 for example `@int` means a managed pointer to a managed box containing an integer.
3146 Copying a managed pointer is a "shallow" operation:
3147 it involves only copying the pointer itself
3148 (as well as any reference-count or GC-barriers required by the managed heap).
3149 Dropping a managed pointer does not necessarily release the box it points to;
3150 the lifecycles of managed boxes are subject to an unspecified garbage collection algorithm.
3152 Owning pointers (`~`)
3153 : These point to owned heap allocations (or "boxes") in the shared, inter-task heap.
3154 Each owned box has a single owning pointer; pointer and pointee retain a 1:1 relationship at all times.
3155 Owning pointers are written `~content`,
3156 for example `~int` means an owning pointer to an owned box containing an integer.
3157 Copying an owned box is a "deep" operation:
3158 it involves allocating a new owned box and copying the contents of the old box into the new box.
3159 Releasing an owning pointer immediately releases its corresponding owned box.
3161 Borrowed pointers (`&`)
3162 : These point to memory _owned by some other value_.
3163 Borrowed pointers arise by (automatic) conversion from owning pointers, managed pointers,
3164 or by applying the borrowing operator `&` to some other value,
3165 including [lvalues, rvalues or temporaries](#lvalues-rvalues-and-temporaries).
3166 Borrowed pointers are written `&content`, or in some cases `&'f content` for some lifetime-variable `f`,
3167 for example `&int` means a borrowed pointer to an integer.
3168 Copying a borrowed pointer is a "shallow" operation:
3169 it involves only copying the pointer itself.
3170 Releasing a borrowed pointer typically has no effect on the value it points to,
3171 with the exception of temporary values,
3172 which are released when the last borrowed pointer to them is released.
3175 : Raw pointers are pointers without safety or liveness guarantees.
3176 Raw pointers are written `*content`,
3177 for example `*int` means a raw pointer to an integer.
3178 Copying or dropping a raw pointer has no effect on the lifecycle of any other value.
3179 Dereferencing a raw pointer or converting it to any other pointer type is an [`unsafe` operation](#unsafe-functions).
3180 Raw pointers are generally discouraged in Rust code;
3181 they exist to support interoperability with foreign code,
3182 and writing performance-critical or low-level functions.
3186 The function type constructor `fn` forms new function types.
3187 A function type consists of a possibly-empty set of function-type modifiers
3188 (such as `unsafe` or `extern`), a sequence of input types and an output type.
3190 An example of a `fn` type:
3193 fn add(x: int, y: int) -> int {
3197 let mut x = add(5,7);
3199 type Binop<'self> = &'self fn(int,int) -> int;
3200 let bo: Binop = add;
3206 Every trait item (see [traits](#traits)) defines a type with the same name as the trait.
3207 This type is called the _object type_ of the trait.
3208 Object types permit "late binding" of methods, dispatched using _virtual method tables_ ("vtables").
3209 Whereas most calls to trait methods are "early bound" (statically resolved) to specific implementations at compile time,
3210 a call to a method on an object type is only resolved to a vtable entry at compile time.
3211 The actual implementation for each vtable entry can vary on an object-by-object basis.
3213 Given a pointer-typed expression `E` of type `&T`, `~T` or `@T`, where `T` implements trait `R`,
3214 casting `E` to the corresponding pointer type `&R`, `~R` or `@R` results in a value of the _object type_ `R`.
3215 This result is represented as a pair of pointers:
3216 the vtable pointer for the `T` implementation of `R`, and the pointer value of `E`.
3218 An example of an object type:
3222 fn to_string(&self) -> ~str;
3225 impl Printable for int {
3226 fn to_string(&self) -> ~str { self.to_str() }
3229 fn print(a: @Printable) {
3230 println(a.to_string());
3234 print(@10 as @Printable);
3238 In this example, the trait `Printable` occurs as an object type in both the type signature of `print`,
3239 and the cast expression in `main`.
3243 Within the body of an item that has type parameter declarations, the names of its type parameters are types:
3246 fn map<A: Clone, B: Clone>(f: |A| -> B, xs: &[A]) -> ~[B] {
3250 let first: B = f(xs[0].clone());
3251 let rest: ~[B] = map(f, xs.slice(1, xs.len()));
3252 return ~[first] + rest;
3256 Here, `first` has type `B`, referring to `map`'s `B` type parameter;
3257 and `rest` has type `~[B]`, a vector type with element type `B`.
3261 The special type `self` has a meaning within methods inside an
3262 impl item. It refers to the type of the implicit `self` argument. For
3267 fn make_string(&self) -> ~str;
3270 impl Printable for ~str {
3271 fn make_string(&self) -> ~str {
3277 `self` refers to the value of type `~str` that is the receiver for a
3278 call to the method `make_string`.
3282 Types in Rust are categorized into kinds, based on various properties of the components of the type.
3286 : Types of this kind are deeply immutable;
3287 they contain no mutable memory locations
3288 directly or indirectly via pointers.
3290 : Types of this kind can be safely sent between tasks.
3291 This kind includes scalars, owning pointers, owned closures, and
3292 structural types containing only other owned types.
3293 All `Send` types are `'static`.
3295 : Types of this kind do not contain any borrowed pointers;
3296 this can be a useful guarantee for code
3297 that breaks borrowing assumptions
3298 using [`unsafe` operations](#unsafe-functions).
3300 : This is not strictly a kind,
3301 but its presence interacts with kinds:
3302 the `Drop` trait provides a single method `drop`
3303 that takes no parameters,
3304 and is run when values of the type are dropped.
3305 Such a method is called a "destructor",
3306 and are always executed in "top-down" order:
3307 a value is completely destroyed
3308 before any of the values it owns run their destructors.
3309 Only `Send` types can implement `Drop`.
3312 : Types with destructors, closure environments,
3313 and various other _non-first-class_ types,
3314 are not copyable at all.
3315 Such types can usually only be accessed through pointers,
3316 or in some cases, moved between mutable locations.
3318 Kinds can be supplied as _bounds_ on type parameters, like traits,
3319 in which case the parameter is constrained to types satisfying that kind.
3321 By default, type parameters do not carry any assumed kind-bounds at all.
3322 When instantiating a type parameter,
3323 the kind bounds on the parameter are checked
3324 to be the same or narrower than the kind
3325 of the type that it is instantiated with.
3327 Sending operations are not part of the Rust language,
3328 but are implemented in the library.
3329 Generic functions that send values
3330 bound the kind of these values to sendable.
3332 # Memory and concurrency models
3334 Rust has a memory model centered around concurrently-executing _tasks_. Thus
3335 its memory model and its concurrency model are best discussed simultaneously,
3336 as parts of each only make sense when considered from the perspective of the
3339 When reading about the memory model, keep in mind that it is partitioned in
3340 order to support tasks; and when reading about tasks, keep in mind that their
3341 isolation and communication mechanisms are only possible due to the ownership
3342 and lifetime semantics of the memory model.
3346 A Rust program's memory consists of a static set of *items*, a set of
3347 [tasks](#tasks) each with its own *stack*, and a *heap*. Immutable portions of
3348 the heap may be shared between tasks, mutable portions may not.
3350 Allocations in the stack consist of *slots*, and allocations in the heap
3353 ### Memory allocation and lifetime
3355 The _items_ of a program are those functions, modules and types
3356 that have their value calculated at compile-time and stored uniquely in the
3357 memory image of the rust process. Items are neither dynamically allocated nor
3360 A task's _stack_ consists of activation frames automatically allocated on
3361 entry to each function as the task executes. A stack allocation is reclaimed
3362 when control leaves the frame containing it.
3364 The _heap_ is a general term that describes two separate sets of boxes:
3365 managed boxes -- which may be subject to garbage collection -- and owned
3366 boxes. The lifetime of an allocation in the heap depends on the lifetime of
3367 the box values pointing to it. Since box values may themselves be passed in
3368 and out of frames, or stored in the heap, heap allocations may outlive the
3369 frame they are allocated within.
3371 ### Memory ownership
3373 A task owns all memory it can *safely* reach through local variables,
3374 as well as managed, owning and borrowed pointers.
3376 When a task sends a value that has the `Send` trait to another task,
3377 it loses ownership of the value sent and can no longer refer to it.
3378 This is statically guaranteed by the combined use of "move semantics",
3379 and the compiler-checked _meaning_ of the `Send` trait:
3380 it is only instantiated for (transitively) sendable kinds of data constructor and pointers,
3381 never including managed or borrowed pointers.
3383 When a stack frame is exited, its local allocations are all released, and its
3384 references to boxes (both managed and owned) are dropped.
3386 A managed box may (in the case of a recursive, mutable managed type) be cyclic;
3387 in this case the release of memory inside the managed structure may be deferred
3388 until task-local garbage collection can reclaim it. Code can ensure no such
3389 delayed deallocation occurs by restricting itself to owned boxes and similar
3390 unmanaged kinds of data.
3392 When a task finishes, its stack is necessarily empty and it therefore has no
3393 references to any boxes; the remainder of its heap is immediately freed.
3397 A task's stack contains slots.
3399 A _slot_ is a component of a stack frame, either a function parameter,
3400 a [temporary](#lvalues-rvalues-and-temporaries), or a local variable.
3402 A _local variable_ (or *stack-local* allocation) holds a value directly,
3403 allocated within the stack's memory. The value is a part of the stack frame.
3405 Local variables are immutable unless declared otherwise like: `let mut x = ...`.
3407 Function parameters are immutable unless declared with `mut`. The
3408 `mut` keyword applies only to the following parameter (so `|mut x, y|`
3409 and `fn f(mut x: ~int, y: ~int)` declare one mutable variable `x` and
3410 one immutable variable `y`).
3412 Methods that take either `self` or `~self` can optionally place them in a
3413 mutable slot by prefixing them with `mut` (similar to regular arguments):
3417 fn change(mut self) -> Self;
3418 fn modify(mut ~self) -> ~Self;
3422 Local variables are not initialized when allocated; the entire frame worth of
3423 local variables are allocated at once, on frame-entry, in an uninitialized
3424 state. Subsequent statements within a function may or may not initialize the
3425 local variables. Local variables can be used only after they have been
3426 initialized; this is enforced by the compiler.
3430 A _box_ is a reference to a heap allocation holding another value. There
3431 are two kinds of boxes: *managed boxes* and *owned boxes*.
3433 A _managed box_ type or value is constructed by the prefix *at* sigil `@`.
3435 An _owned box_ type or value is constructed by the prefix *tilde* sigil `~`.
3437 Multiple managed box values can point to the same heap allocation; copying a
3438 managed box value makes a shallow copy of the pointer (optionally incrementing
3439 a reference count, if the managed box is implemented through
3440 reference-counting).
3442 Owned box values exist in 1:1 correspondence with their heap allocation.
3444 An example of constructing one managed box type and value, and one owned box
3452 Some operations (such as field selection) implicitly dereference boxes. An
3453 example of an _implicit dereference_ operation performed on box values:
3456 struct Foo { y: int }
3457 let x = @Foo{y: 10};
3461 Other operations act on box values as single-word-sized address values. For
3462 these operations, to access the value held in the box requires an explicit
3463 dereference of the box value. Explicitly dereferencing a box is indicated with
3464 the unary *star* operator `*`. Examples of such _explicit dereference_
3467 * copying box values (`x = y`)
3468 * passing box values to functions (`f(x,y)`)
3470 An example of an explicit-dereference operation performed on box values:
3473 fn takes_boxed(b: @int) {
3476 fn takes_unboxed(b: int) {
3488 An executing Rust program consists of a tree of tasks.
3489 A Rust _task_ consists of an entry function, a stack,
3490 a set of outgoing communication channels and incoming communication ports,
3491 and ownership of some portion of the heap of a single operating-system process.
3492 (We expect that many programs will not use channels and ports directly,
3493 but will instead use higher-level abstractions provided in standard libraries,
3496 Multiple Rust tasks may coexist in a single operating-system process.
3497 The runtime scheduler maps tasks to a certain number of operating-system threads.
3498 By default, the scheduler chooses the number of threads based on
3499 the number of concurrent physical CPUs detected at startup.
3500 It's also possible to override this choice at runtime.
3501 When the number of tasks exceeds the number of threads -- which is likely --
3502 the scheduler multiplexes the tasks onto threads.^[
3503 This is an M:N scheduler,
3504 which is known to give suboptimal results for CPU-bound concurrency problems.
3505 In such cases, running with the same number of threads and tasks can yield better results.
3506 Rust has M:N scheduling in order to support very large numbers of tasks
3507 in contexts where threads are too resource-intensive to use in large number.
3508 The cost of threads varies substantially per operating system, and is sometimes quite low,
3509 so this flexibility is not always worth exploiting.]
3511 ### Communication between tasks
3513 Rust tasks are isolated and generally unable to interfere with one another's memory directly,
3514 except through [`unsafe` code](#unsafe-functions).
3515 All contact between tasks is mediated by safe forms of ownership transfer,
3516 and data races on memory are prohibited by the type system.
3518 Inter-task communication and co-ordination facilities are provided in the standard library.
3521 - synchronous and asynchronous communication channels with various communication topologies
3522 - read-only and read-write shared variables with various safe mutual exclusion patterns
3523 - simple locks and semaphores
3525 When such facilities carry values, the values are restricted to the [`Send` type-kind](#type-kinds).
3526 Restricting communication interfaces to this kind ensures that no borrowed or managed pointers move between tasks.
3527 Thus access to an entire data structure can be mediated through its owning "root" value;
3528 no further locking or copying is required to avoid data races within the substructure of such a value.
3532 The _lifecycle_ of a task consists of a finite set of states and events
3533 that cause transitions between the states. The lifecycle states of a task are:
3540 A task begins its lifecycle -- once it has been spawned -- in the *running*
3541 state. In this state it executes the statements of its entry function, and any
3542 functions called by the entry function.
3544 A task may transition from the *running* state to the *blocked*
3545 state any time it makes a blocking communication call. When the
3546 call can be completed -- when a message arrives at a sender, or a
3547 buffer opens to receive a message -- then the blocked task will
3548 unblock and transition back to *running*.
3550 A task may transition to the *failing* state at any time, due being
3551 killed by some external event or internally, from the evaluation of a
3552 `fail!()` macro. Once *failing*, a task unwinds its stack and
3553 transitions to the *dead* state. Unwinding the stack of a task is done by
3554 the task itself, on its own control stack. If a value with a destructor is
3555 freed during unwinding, the code for the destructor is run, also on the task's
3556 control stack. Running the destructor code causes a temporary transition to a
3557 *running* state, and allows the destructor code to cause any subsequent
3558 state transitions. The original task of unwinding and failing thereby may
3559 suspend temporarily, and may involve (recursive) unwinding of the stack of a
3560 failed destructor. Nonetheless, the outermost unwinding activity will continue
3561 until the stack is unwound and the task transitions to the *dead*
3562 state. There is no way to "recover" from task failure. Once a task has
3563 temporarily suspended its unwinding in the *failing* state, failure
3564 occurring from within this destructor results in *hard* failure. The
3565 unwinding procedure of hard failure frees resources but does not execute
3566 destructors. The original (soft) failure is still resumed at the point where
3567 it was temporarily suspended.
3569 A task in the *dead* state cannot transition to other states; it exists
3570 only to have its termination status inspected by other tasks, and/or to await
3571 reclamation when the last reference to it drops.
3575 The currently scheduled task is given a finite *time slice* in which to
3576 execute, after which it is *descheduled* at a loop-edge or similar
3577 preemption point, and another task within is scheduled, pseudo-randomly.
3579 An executing task can yield control at any time, by making a library call to
3580 `std::task::yield`, which deschedules it immediately. Entering any other
3581 non-executing state (blocked, dead) similarly deschedules the task.
3583 # Runtime services, linkage and debugging
3585 The Rust _runtime_ is a relatively compact collection of C++ and Rust code
3586 that provides fundamental services and datatypes to all Rust tasks at
3587 run-time. It is smaller and simpler than many modern language runtimes. It is
3588 tightly integrated into the language's execution model of memory, tasks,
3589 communication and logging.
3591 > **Note:** The runtime library will merge with the `std` library in future versions of Rust.
3593 ### Memory allocation
3595 The runtime memory-management system is based on a _service-provider interface_,
3596 through which the runtime requests blocks of memory from its environment
3597 and releases them back to its environment when they are no longer needed.
3598 The default implementation of the service-provider interface
3599 consists of the C runtime functions `malloc` and `free`.
3601 The runtime memory-management system, in turn, supplies Rust tasks with
3602 facilities for allocating releasing stacks, as well as allocating and freeing
3607 The runtime provides C and Rust code to assist with various built-in types,
3608 such as vectors, strings, and the low level communication system (ports,
3611 Support for other built-in types such as simple types, tuples, records, and
3612 enums is open-coded by the Rust compiler.
3614 ### Task scheduling and communication
3616 The runtime provides code to manage inter-task communication. This includes
3617 the system of task-lifecycle state transitions depending on the contents of
3618 queues, as well as code to copy values between queues and their recipients and
3619 to serialize values for transmission over operating-system inter-process
3620 communication facilities.
3624 The runtime contains a system for directing [logging
3625 expressions](#log-expressions) to a logging console and/or internal logging
3626 buffers. Logging can be enabled per module.
3628 Logging output is enabled by setting the `RUST_LOG` environment
3629 variable. `RUST_LOG` accepts a logging specification made up of a
3630 comma-separated list of paths, with optional log levels. For each
3631 module containing log expressions, if `RUST_LOG` contains the path to
3632 that module or a parent of that module, then logs of the appropriate
3633 level will be output to the console.
3635 The path to a module consists of the crate name, any parent modules,
3636 then the module itself, all separated by double colons (`::`). The
3637 optional log level can be appended to the module path with an equals
3638 sign (`=`) followed by the log level, from 1 to 4, inclusive. Level 1
3639 is the error level, 2 is warning, 3 info, and 4 debug. You can also
3640 use the symbolic constants `error`, `warn`, `info`, and `debug`. Any
3641 logs less than or equal to the specified level will be output. If not
3642 specified then log level 4 is assumed. Debug messages can be omitted
3643 by passing `--cfg ndebug` to `rustc`.
3645 As an example, to see all the logs generated by the compiler, you would set
3646 `RUST_LOG` to `rustc`, which is the crate name (as specified in its `link`
3647 [attribute](#attributes)). To narrow down the logs to just crate resolution,
3648 you would set it to `rustc::metadata::creader`. To see just error logging
3651 Note that when compiling source files that don't specify a
3652 crate name the crate is given a default name that matches the source file,
3653 with the extension removed. In that case, to turn on logging for a program
3654 compiled from, e.g. `helloworld.rs`, `RUST_LOG` should be set to `helloworld`.
3656 As a convenience, the logging spec can also be set to a special pseudo-crate,
3657 `::help`. In this case, when the application starts, the runtime will
3658 simply output a list of loaded modules containing log expressions, then exit.
3660 The Rust runtime itself generates logging information. The runtime's logs are
3661 generated for a number of artificial modules in the `::rt` pseudo-crate,
3662 and can be enabled just like the logs for any standard module. The full list
3663 of runtime logging modules follows.
3665 * `::rt::mem` Memory management
3666 * `::rt::comm` Messaging and task communication
3667 * `::rt::task` Task management
3668 * `::rt::dom` Task scheduling
3669 * `::rt::trace` Unused
3670 * `::rt::cache` Type descriptor cache
3671 * `::rt::upcall` Compiler-generated runtime calls
3672 * `::rt::timer` The scheduler timer
3673 * `::rt::gc` Garbage collection
3674 * `::rt::stdlib` Functions used directly by the standard library
3675 * `::rt::kern` The runtime kernel
3676 * `::rt::backtrace` Log a backtrace on task failure
3677 * `::rt::callback` Unused
3679 #### Logging Expressions
3681 Rust provides several macros to log information. Here's a simple Rust program
3682 that demonstrates all four of them:
3686 error!("This is an error log")
3687 warn!("This is a warn log")
3688 info!("this is an info log")
3689 debug!("This is a debug log")
3693 These four log levels correspond to levels 1-4, as controlled by `RUST_LOG`:
3696 $ RUST_LOG=rust=3 ./rust
3697 This is an error log
3702 # Appendix: Rationales and design tradeoffs
3706 # Appendix: Influences and further references
3710 > The essential problem that must be solved in making a fault-tolerant
3711 > software system is therefore that of fault-isolation. Different programmers
3712 > will write different modules, some modules will be correct, others will have
3713 > errors. We do not want the errors in one module to adversely affect the
3714 > behaviour of a module which does not have any errors.
3716 > — Joe Armstrong
3718 > In our approach, all data is private to some process, and processes can
3719 > only communicate through communications channels. *Security*, as used
3720 > in this paper, is the property which guarantees that processes in a system
3721 > cannot affect each other except by explicit communication.
3723 > When security is absent, nothing which can be proven about a single module
3724 > in isolation can be guaranteed to hold when that module is embedded in a
3727 > — Robert Strom and Shaula Yemini
3729 > Concurrent and applicative programming complement each other. The
3730 > ability to send messages on channels provides I/O without side effects,
3731 > while the avoidance of shared data helps keep concurrent processes from
3736 Rust is not a particularly original language. It may however appear unusual
3737 by contemporary standards, as its design elements are drawn from a number of
3738 "historical" languages that have, with a few exceptions, fallen out of
3739 favour. Five prominent lineages contribute the most, though their influences
3740 have come and gone during the course of Rust's development:
3742 * The NIL (1981) and Hermes (1990) family. These languages were developed by
3743 Robert Strom, Shaula Yemini, David Bacon and others in their group at IBM
3744 Watson Research Center (Yorktown Heights, NY, USA).
3746 * The Erlang (1987) language, developed by Joe Armstrong, Robert Virding, Claes
3747 Wikström, Mike Williams and others in their group at the Ericsson Computer
3748 Science Laboratory (Älvsjö, Stockholm, Sweden) .
3750 * The Sather (1990) language, developed by Stephen Omohundro, Chu-Cheow Lim,
3751 Heinz Schmidt and others in their group at The International Computer
3752 Science Institute of the University of California, Berkeley (Berkeley, CA,
3755 * The Newsqueak (1988), Alef (1995), and Limbo (1996) family. These
3756 languages were developed by Rob Pike, Phil Winterbottom, Sean Dorward and
3757 others in their group at Bell Labs Computing Sciences Research Center
3758 (Murray Hill, NJ, USA).
3760 * The Napier (1985) and Napier88 (1988) family. These languages were
3761 developed by Malcolm Atkinson, Ron Morrison and others in their group at
3762 the University of St. Andrews (St. Andrews, Fife, UK).
3764 Additional specific influences can be seen from the following languages:
3766 * The structural algebraic types and compilation manager of SML.
3767 * The attribute and assembly systems of C#.
3768 * The references and deterministic destructor system of C++.
3769 * The memory region systems of the ML Kit and Cyclone.
3770 * The typeclass system of Haskell.
3771 * The lexical identifier rule of Python.
3772 * The block syntax of Ruby.