src/libsyntax/parse/token.rs

   1 // Copyright 2012-2013 The Rust Project Developers. See the COPYRIGHT
   2 // file at the top-level directory of this distribution and at
   3 // http://rust-lang.org/COPYRIGHT.
   4 //
   5 // Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
   6 // http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
   7 // <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
   8 // option. This file may not be copied, modified, or distributed
   9 // except according to those terms.
  10
  11 pub use self::BinOpToken::*;
  12 pub use self::Nonterminal::*;
  13 pub use self::DelimToken::*;
  14 pub use self::IdentStyle::*;
  15 pub use self::Lit::*;
  16 pub use self::Token::*;
  17
  18 use ast;
  19 use ext::mtwt;
  20 use ptr::P;
  21 use util::interner::{RcStr, StrInterner};
  22 use util::interner;
  23
  24 use serialize::{Decodable, Decoder, Encodable, Encoder};
  25 use std::fmt;
  26 use std::ops::Deref;
  27 use std::rc::Rc;
  28
  29 #[allow(non_camel_case_types)]
  30 #[derive(Clone, RustcEncodable, RustcDecodable, PartialEq, Eq, Hash, Debug, Copy)]
  31 pub enum BinOpToken {
  32     Plus,
  33     Minus,
  34     Star,
  35     Slash,
  36     Percent,
  37     Caret,
  38     And,
  39     Or,
  40     Shl,
  41     Shr,
  42 }
  43
  44 /// A delimiter token
  45 #[derive(Clone, RustcEncodable, RustcDecodable, PartialEq, Eq, Hash, Debug, Copy)]
  46 pub enum DelimToken {
  47     /// A round parenthesis: `(` or `)`
  48     Paren,
  49     /// A square bracket: `[` or `]`
  50     Bracket,
  51     /// A curly brace: `{` or `}`
  52     Brace,
  53 }
  54
  55 #[derive(Clone, RustcEncodable, RustcDecodable, PartialEq, Eq, Hash, Debug, Copy)]
  56 pub enum IdentStyle {
  57     /// `::` follows the identifier with no whitespace in-between.
  58     ModName,
  59     Plain,
  60 }
  61
  62 #[derive(Clone, RustcEncodable, RustcDecodable, PartialEq, Eq, Hash, Debug, Copy)]
  63 pub enum SpecialMacroVar {
  64     /// `$crate` will be filled in with the name of the crate a macro was
  65     /// imported from, if any.
  66     CrateMacroVar,
  67 }
  68
  69 impl SpecialMacroVar {
  70     pub fn as_str(self) -> &'static str {
  71         match self {
  72             SpecialMacroVar::CrateMacroVar => "crate",
  73         }
  74     }
  75 }
  76
  77 #[derive(Clone, RustcEncodable, RustcDecodable, PartialEq, Eq, Hash, Debug, Copy)]
  78 pub enum Lit {
  79     Byte(ast::Name),
  80     Char(ast::Name),
  81     Integer(ast::Name),
  82     Float(ast::Name),
  83     Str_(ast::Name),
  84     StrRaw(ast::Name, usize), /* raw str delimited by n hash symbols */
  85     ByteStr(ast::Name),
  86     ByteStrRaw(ast::Name, usize), /* raw byte str delimited by n hash symbols */
  87 }
  88
  89 impl Lit {
  90     pub fn short_name(&self) -> &'static str {
  91         match *self {
  92             Byte(_) => "byte",
  93             Char(_) => "char",
  94             Integer(_) => "integer",
  95             Float(_) => "float",
  96             Str_(_) | StrRaw(..) => "string",
  97             ByteStr(_) | ByteStrRaw(..) => "byte string"
  98         }
  99     }
 100 }
 101
 102 #[allow(non_camel_case_types)]
 103 #[derive(Clone, RustcEncodable, RustcDecodable, PartialEq, Eq, Hash, Debug)]
 104 pub enum Token {
 105     /* Expression-operator symbols. */
 106     Eq,
 107     Lt,
 108     Le,
 109     EqEq,
 110     Ne,
 111     Ge,
 112     Gt,
 113     AndAnd,
 114     OrOr,
 115     Not,
 116     Tilde,
 117     BinOp(BinOpToken),
 118     BinOpEq(BinOpToken),
 119
 120     /* Structural symbols */
 121     At,
 122     Dot,
 123     DotDot,
 124     DotDotDot,
 125     Comma,
 126     Semi,
 127     Colon,
 128     ModSep,
 129     RArrow,
 130     LArrow,
 131     FatArrow,
 132     Pound,
 133     Dollar,
 134     Question,
 135     /// An opening delimiter, eg. `{`
 136     OpenDelim(DelimToken),
 137     /// A closing delimiter, eg. `}`
 138     CloseDelim(DelimToken),
 139
 140     /* Literals */
 141     Literal(Lit, Option<ast::Name>),
 142
 143     /* Name components */
 144     Ident(ast::Ident, IdentStyle),
 145     Underscore,
 146     Lifetime(ast::Ident),
 147
 148     /* For interpolation */
 149     Interpolated(Nonterminal),
 150     // Can be expanded into several tokens.
 151     /// Doc comment
 152     DocComment(ast::Name),
 153     // In left-hand-sides of MBE macros:
 154     /// Parse a nonterminal (name to bind, name of NT, styles of their idents)
 155     MatchNt(ast::Ident, ast::Ident, IdentStyle, IdentStyle),
 156     // In right-hand-sides of MBE macros:
 157     /// A syntactic variable that will be filled in by macro expansion.
 158     SubstNt(ast::Ident, IdentStyle),
 159     /// A macro variable with special meaning.
 160     SpecialVarNt(SpecialMacroVar),
 161
 162     // Junk. These carry no data because we don't really care about the data
 163     // they *would* carry, and don't really want to allocate a new ident for
 164     // them. Instead, users could extract that from the associated span.
 165
 166     /// Whitespace
 167     Whitespace,
 168     /// Comment
 169     Comment,
 170     Shebang(ast::Name),
 171
 172     Eof,
 173 }
 174
 175 impl Token {
 176     /// Returns `true` if the token starts with '>'.
 177     pub fn is_like_gt(&self) -> bool {
 178         match *self {
 179             BinOp(Shr) | BinOpEq(Shr) | Gt | Ge => true,
 180             _ => false,
 181         }
 182     }
 183
 184     /// Returns `true` if the token can appear at the start of an expression.
 185     pub fn can_begin_expr(&self) -> bool {
 186         match *self {
 187             OpenDelim(_)                => true,
 188             Ident(_, _)                 => true,
 189             Underscore                  => true,
 190             Tilde                       => true,
 191             Literal(_, _)               => true,
 192             Not                         => true,
 193             BinOp(Minus)                => true,
 194             BinOp(Star)                 => true,
 195             BinOp(And)                  => true,
 196             BinOp(Or)                   => true, // in lambda syntax
 197             OrOr                        => true, // in lambda syntax
 198             AndAnd                      => true, // double borrow
 199             DotDot                      => true, // range notation
 200             ModSep                      => true,
 201             Interpolated(NtExpr(..))    => true,
 202             Interpolated(NtIdent(..))   => true,
 203             Interpolated(NtBlock(..))   => true,
 204             Interpolated(NtPath(..))    => true,
 205             Pound                       => true, // for expression attributes
 206             _                           => false,
 207         }
 208     }
 209
 210     /// Returns `true` if the token is any literal
 211     pub fn is_lit(&self) -> bool {
 212         match *self {
 213             Literal(_, _) => true,
 214             _          => false,
 215         }
 216     }
 217
 218     /// Returns `true` if the token is an identifier.
 219     pub fn is_ident(&self) -> bool {
 220         match *self {
 221             Ident(_, _) => true,
 222             _           => false,
 223         }
 224     }
 225
 226     /// Returns `true` if the token is an interpolated path.
 227     pub fn is_path(&self) -> bool {
 228         match *self {
 229             Interpolated(NtPath(..))    => true,
 230             _                           => false,
 231         }
 232     }
 233
 234     /// Returns `true` if the token is a path that is not followed by a `::`
 235     /// token.
 236     #[allow(non_upper_case_globals)]
 237     pub fn is_plain_ident(&self) -> bool {
 238         match *self {
 239             Ident(_, Plain) => true,
 240             _               => false,
 241         }
 242     }
 243
 244     /// Returns `true` if the token is a lifetime.
 245     pub fn is_lifetime(&self) -> bool {
 246         match *self {
 247             Lifetime(..) => true,
 248             _            => false,
 249         }
 250     }
 251
 252     /// Returns `true` if the token is either the `mut` or `const` keyword.
 253     pub fn is_mutability(&self) -> bool {
 254         self.is_keyword(keywords::Mut) ||
 255         self.is_keyword(keywords::Const)
 256     }
 257
 258     /// Maps a token to its corresponding binary operator.
 259     pub fn to_binop(&self) -> Option<ast::BinOp_> {
 260         match *self {
 261             BinOp(Star)     => Some(ast::BiMul),
 262             BinOp(Slash)    => Some(ast::BiDiv),
 263             BinOp(Percent)  => Some(ast::BiRem),
 264             BinOp(Plus)     => Some(ast::BiAdd),
 265             BinOp(Minus)    => Some(ast::BiSub),
 266             BinOp(Shl)      => Some(ast::BiShl),
 267             BinOp(Shr)      => Some(ast::BiShr),
 268             BinOp(And)      => Some(ast::BiBitAnd),
 269             BinOp(Caret)    => Some(ast::BiBitXor),
 270             BinOp(Or)       => Some(ast::BiBitOr),
 271             Lt              => Some(ast::BiLt),
 272             Le              => Some(ast::BiLe),
 273             Ge              => Some(ast::BiGe),
 274             Gt              => Some(ast::BiGt),
 275             EqEq            => Some(ast::BiEq),
 276             Ne              => Some(ast::BiNe),
 277             AndAnd          => Some(ast::BiAnd),
 278             OrOr            => Some(ast::BiOr),
 279             _               => None,
 280         }
 281     }
 282
 283     /// Returns `true` if the token is a given keyword, `kw`.
 284     #[allow(non_upper_case_globals)]
 285     pub fn is_keyword(&self, kw: keywords::Keyword) -> bool {
 286         match *self {
 287             Ident(sid, Plain) => kw.to_name() == sid.name,
 288             _                      => false,
 289         }
 290     }
 291
 292     pub fn is_keyword_allow_following_colon(&self, kw: keywords::Keyword) -> bool {
 293         match *self {
 294             Ident(sid, _) => { kw.to_name() == sid.name }
 295             _ => { false }
 296         }
 297     }
 298
 299     /// Returns `true` if the token is either a special identifier, or a strict
 300     /// or reserved keyword.
 301     #[allow(non_upper_case_globals)]
 302     pub fn is_any_keyword(&self) -> bool {
 303         match *self {
 304             Ident(sid, Plain) => {
 305                 let n = sid.name;
 306
 307                    n == SELF_KEYWORD_NAME
 308                 || n == STATIC_KEYWORD_NAME
 309                 || n == SUPER_KEYWORD_NAME
 310                 || n == SELF_TYPE_KEYWORD_NAME
 311                 || STRICT_KEYWORD_START <= n
 312                 && n <= RESERVED_KEYWORD_FINAL
 313             },
 314             _ => false
 315         }
 316     }
 317
 318     /// Returns `true` if the token may not appear as an identifier.
 319     #[allow(non_upper_case_globals)]
 320     pub fn is_strict_keyword(&self) -> bool {
 321         match *self {
 322             Ident(sid, Plain) => {
 323                 let n = sid.name;
 324
 325                    n == SELF_KEYWORD_NAME
 326                 || n == STATIC_KEYWORD_NAME
 327                 || n == SUPER_KEYWORD_NAME
 328                 || n == SELF_TYPE_KEYWORD_NAME
 329                 || STRICT_KEYWORD_START <= n
 330                 && n <= STRICT_KEYWORD_FINAL
 331             },
 332             Ident(sid, ModName) => {
 333                 let n = sid.name;
 334
 335                    n != SELF_KEYWORD_NAME
 336                 && n != SUPER_KEYWORD_NAME
 337                 && STRICT_KEYWORD_START <= n
 338                 && n <= STRICT_KEYWORD_FINAL
 339             }
 340             _ => false,
 341         }
 342     }
 343
 344     /// Returns `true` if the token is a keyword that has been reserved for
 345     /// possible future use.
 346     #[allow(non_upper_case_globals)]
 347     pub fn is_reserved_keyword(&self) -> bool {
 348         match *self {
 349             Ident(sid, Plain) => {
 350                 let n = sid.name;
 351
 352                    RESERVED_KEYWORD_START <= n
 353                 && n <= RESERVED_KEYWORD_FINAL
 354             },
 355             _ => false,
 356         }
 357     }
 358
 359     /// Hygienic identifier equality comparison.
 360     ///
 361     /// See `styntax::ext::mtwt`.
 362     pub fn mtwt_eq(&self, other : &Token) -> bool {
 363         match (self, other) {
 364             (&Ident(id1,_), &Ident(id2,_)) | (&Lifetime(id1), &Lifetime(id2)) =>
 365                 mtwt::resolve(id1) == mtwt::resolve(id2),
 366             _ => *self == *other
 367         }
 368     }
 369 }
 370
 371 #[derive(Clone, RustcEncodable, RustcDecodable, PartialEq, Eq, Hash)]
 372 /// For interpolation during macro expansion.
 373 pub enum Nonterminal {
 374     NtItem(P<ast::Item>),
 375     NtBlock(P<ast::Block>),
 376     NtStmt(P<ast::Stmt>),
 377     NtPat(P<ast::Pat>),
 378     NtExpr(P<ast::Expr>),
 379     NtTy(P<ast::Ty>),
 380     NtIdent(Box<ast::SpannedIdent>, IdentStyle),
 381     /// Stuff inside brackets for attributes
 382     NtMeta(P<ast::MetaItem>),
 383     NtPath(Box<ast::Path>),
 384     NtTT(P<ast::TokenTree>), // needs P'ed to break a circularity
 385     // These are not exposed to macros, but are used by quasiquote.
 386     NtArm(ast::Arm),
 387     NtImplItem(P<ast::ImplItem>),
 388     NtTraitItem(P<ast::TraitItem>),
 389     NtGenerics(ast::Generics),
 390     NtWhereClause(ast::WhereClause),
 391     NtArg(ast::Arg),
 392 }
 393
 394 impl fmt::Debug for Nonterminal {
 395     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
 396         match *self {
 397             NtItem(..) => f.pad("NtItem(..)"),
 398             NtBlock(..) => f.pad("NtBlock(..)"),
 399             NtStmt(..) => f.pad("NtStmt(..)"),
 400             NtPat(..) => f.pad("NtPat(..)"),
 401             NtExpr(..) => f.pad("NtExpr(..)"),
 402             NtTy(..) => f.pad("NtTy(..)"),
 403             NtIdent(..) => f.pad("NtIdent(..)"),
 404             NtMeta(..) => f.pad("NtMeta(..)"),
 405             NtPath(..) => f.pad("NtPath(..)"),
 406             NtTT(..) => f.pad("NtTT(..)"),
 407             NtArm(..) => f.pad("NtArm(..)"),
 408             NtImplItem(..) => f.pad("NtImplItem(..)"),
 409             NtTraitItem(..) => f.pad("NtTraitItem(..)"),
 410             NtGenerics(..) => f.pad("NtGenerics(..)"),
 411             NtWhereClause(..) => f.pad("NtWhereClause(..)"),
 412             NtArg(..) => f.pad("NtArg(..)"),
 413         }
 414     }
 415 }
 416
 417
 418 // Get the first "argument"
 419 macro_rules! first {
 420     ( $first:expr, $( $remainder:expr, )* ) => ( $first )
 421 }
 422
 423 // Get the last "argument" (has to be done recursively to avoid phoney local ambiguity error)
 424 macro_rules! last {
 425     ( $first:expr, $( $remainder:expr, )+ ) => ( last!( $( $remainder, )+ ) );
 426     ( $first:expr, ) => ( $first )
 427 }
 428
 429 // In this macro, there is the requirement that the name (the number) must be monotonically
 430 // increasing by one in the special identifiers, starting at 0; the same holds for the keywords,
 431 // except starting from the next number instead of zero, and with the additional exception that
 432 // special identifiers are *also* allowed (they are deduplicated in the important place, the
 433 // interner), an exception which is demonstrated by "static" and "self".
 434 macro_rules! declare_special_idents_and_keywords {(
 435     // So now, in these rules, why is each definition parenthesised?
 436     // Answer: otherwise we get a spurious local ambiguity bug on the "}"
 437     pub mod special_idents {
 438         $( ($si_name:expr, $si_static:ident, $si_str:expr); )*
 439     }
 440
 441     pub mod keywords {
 442         'strict:
 443         $( ($sk_name:expr, $sk_variant:ident, $sk_str:expr); )*
 444         'reserved:
 445         $( ($rk_name:expr, $rk_variant:ident, $rk_str:expr); )*
 446     }
 447 ) => {
 448     const STRICT_KEYWORD_START: ast::Name = first!($( ast::Name($sk_name), )*);
 449     const STRICT_KEYWORD_FINAL: ast::Name = last!($( ast::Name($sk_name), )*);
 450     const RESERVED_KEYWORD_START: ast::Name = first!($( ast::Name($rk_name), )*);
 451     const RESERVED_KEYWORD_FINAL: ast::Name = last!($( ast::Name($rk_name), )*);
 452
 453     pub mod special_idents {
 454         use ast;
 455         $(
 456             #[allow(non_upper_case_globals)]
 457             pub const $si_static: ast::Ident = ast::Ident {
 458                 name: ast::Name($si_name),
 459                 ctxt: ast::EMPTY_CTXT,
 460             };
 461          )*
 462     }
 463
 464     pub mod special_names {
 465         use ast;
 466         $(
 467             #[allow(non_upper_case_globals)]
 468             pub const $si_static: ast::Name = ast::Name($si_name);
 469         )*
 470     }
 471
 472     /// All the valid words that have meaning in the Rust language.
 473     ///
 474     /// Rust keywords are either 'strict' or 'reserved'.  Strict keywords may not
 475     /// appear as identifiers at all. Reserved keywords are not used anywhere in
 476     /// the language and may not appear as identifiers.
 477     pub mod keywords {
 478         pub use self::Keyword::*;
 479         use ast;
 480
 481         #[derive(Copy, Clone, PartialEq, Eq)]
 482         pub enum Keyword {
 483             $( $sk_variant, )*
 484             $( $rk_variant, )*
 485         }
 486
 487         impl Keyword {
 488             pub fn to_name(&self) -> ast::Name {
 489                 match *self {
 490                     $( $sk_variant => ast::Name($sk_name), )*
 491                     $( $rk_variant => ast::Name($rk_name), )*
 492                 }
 493             }
 494         }
 495     }
 496
 497     fn mk_fresh_ident_interner() -> IdentInterner {
 498         let mut init_vec = Vec::new();
 499         $(init_vec.push($si_str);)*
 500         $(init_vec.push($sk_str);)*
 501         $(init_vec.push($rk_str);)*
 502         interner::StrInterner::prefill(&init_vec[..])
 503     }
 504 }}
 505
 506 // If the special idents get renumbered, remember to modify these two as appropriate
 507 pub const SELF_KEYWORD_NAME: ast::Name = ast::Name(SELF_KEYWORD_NAME_NUM);
 508 const STATIC_KEYWORD_NAME: ast::Name = ast::Name(STATIC_KEYWORD_NAME_NUM);
 509 const SUPER_KEYWORD_NAME: ast::Name = ast::Name(SUPER_KEYWORD_NAME_NUM);
 510 const SELF_TYPE_KEYWORD_NAME: ast::Name = ast::Name(SELF_TYPE_KEYWORD_NAME_NUM);
 511
 512 pub const SELF_KEYWORD_NAME_NUM: u32 = 1;
 513 const STATIC_KEYWORD_NAME_NUM: u32 = 2;
 514 const SUPER_KEYWORD_NAME_NUM: u32 = 3;
 515 const SELF_TYPE_KEYWORD_NAME_NUM: u32 = 10;
 516
 517 // NB: leaving holes in the ident table is bad! a different ident will get
 518 // interned with the id from the hole, but it will be between the min and max
 519 // of the reserved words, and thus tagged as "reserved".
 520
 521 declare_special_idents_and_keywords! {
 522     pub mod special_idents {
 523         // These ones are statics
 524         (0,                          invalid,                "");
 525         (super::SELF_KEYWORD_NAME_NUM,   self_,              "self");
 526         (super::STATIC_KEYWORD_NAME_NUM, statik,             "static");
 527         (super::SUPER_KEYWORD_NAME_NUM, super_,              "super");
 528         (4,                          static_lifetime,        "'static");
 529
 530         // for matcher NTs
 531         (5,                          tt,                     "tt");
 532         (6,                          matchers,               "matchers");
 533
 534         // outside of libsyntax
 535         (7,                          clownshoe_abi,          "__rust_abi");
 536         (8,                          opaque,                 "<opaque>");
 537         (9,                          unnamed_field,          "<unnamed_field>");
 538         (super::SELF_TYPE_KEYWORD_NAME_NUM, type_self,       "Self");
 539         (11,                         prelude_import,         "prelude_import");
 540     }
 541
 542     pub mod keywords {
 543         // These ones are variants of the Keyword enum
 544
 545         'strict:
 546         (12,                         As,         "as");
 547         (13,                         Break,      "break");
 548         (14,                         Crate,      "crate");
 549         (15,                         Else,       "else");
 550         (16,                         Enum,       "enum");
 551         (17,                         Extern,     "extern");
 552         (18,                         False,      "false");
 553         (19,                         Fn,         "fn");
 554         (20,                         For,        "for");
 555         (21,                         If,         "if");
 556         (22,                         Impl,       "impl");
 557         (23,                         In,         "in");
 558         (24,                         Let,        "let");
 559         (25,                         Loop,       "loop");
 560         (26,                         Match,      "match");
 561         (27,                         Mod,        "mod");
 562         (28,                         Move,       "move");
 563         (29,                         Mut,        "mut");
 564         (30,                         Pub,        "pub");
 565         (31,                         Ref,        "ref");
 566         (32,                         Return,     "return");
 567         // Static and Self are also special idents (prefill de-dupes)
 568         (super::STATIC_KEYWORD_NAME_NUM, Static, "static");
 569         (super::SELF_KEYWORD_NAME_NUM, SelfValue, "self");
 570         (super::SELF_TYPE_KEYWORD_NAME_NUM, SelfType, "Self");
 571         (33,                         Struct,     "struct");
 572         (super::SUPER_KEYWORD_NAME_NUM, Super,   "super");
 573         (34,                         True,       "true");
 574         (35,                         Trait,      "trait");
 575         (36,                         Type,       "type");
 576         (37,                         Unsafe,     "unsafe");
 577         (38,                         Use,        "use");
 578         (39,                         While,      "while");
 579         (40,                         Continue,   "continue");
 580         (41,                         Box,        "box");
 581         (42,                         Const,      "const");
 582         (43,                         Where,      "where");
 583         'reserved:
 584         (44,                         Virtual,    "virtual");
 585         (45,                         Proc,       "proc");
 586         (46,                         Alignof,    "alignof");
 587         (47,                         Become,     "become");
 588         (48,                         Offsetof,   "offsetof");
 589         (49,                         Priv,       "priv");
 590         (50,                         Pure,       "pure");
 591         (51,                         Sizeof,     "sizeof");
 592         (52,                         Typeof,     "typeof");
 593         (53,                         Unsized,    "unsized");
 594         (54,                         Yield,      "yield");
 595         (55,                         Do,         "do");
 596         (56,                         Abstract,   "abstract");
 597         (57,                         Final,      "final");
 598         (58,                         Override,   "override");
 599         (59,                         Macro,      "macro");
 600     }
 601 }
 602
 603 // looks like we can get rid of this completely...
 604 pub type IdentInterner = StrInterner;
 605
 606 // if an interner exists in TLS, return it. Otherwise, prepare a
 607 // fresh one.
 608 // FIXME(eddyb) #8726 This should probably use a thread-local reference.
 609 pub fn get_ident_interner() -> Rc<IdentInterner> {
 610     thread_local!(static KEY: Rc<::parse::token::IdentInterner> = {
 611         Rc::new(mk_fresh_ident_interner())
 612     });
 613     KEY.with(|k| k.clone())
 614 }
 615
 616 /// Reset the ident interner to its initial state.
 617 pub fn reset_ident_interner() {
 618     let interner = get_ident_interner();
 619     interner.reset(mk_fresh_ident_interner());
 620 }
 621
 622 /// Represents a string stored in the thread-local interner. Because the
 623 /// interner lives for the life of the thread, this can be safely treated as an
 624 /// immortal string, as long as it never crosses between threads.
 625 ///
 626 /// FIXME(pcwalton): You must be careful about what you do in the destructors
 627 /// of objects stored in TLS, because they may run after the interner is
 628 /// destroyed. In particular, they must not access string contents. This can
 629 /// be fixed in the future by just leaking all strings until thread death
 630 /// somehow.
 631 #[derive(Clone, PartialEq, Hash, PartialOrd, Eq, Ord)]
 632 pub struct InternedString {
 633     string: RcStr,
 634 }
 635
 636 impl InternedString {
 637     #[inline]
 638     pub fn new(string: &'static str) -> InternedString {
 639         InternedString {
 640             string: RcStr::new(string),
 641         }
 642     }
 643
 644     #[inline]
 645     fn new_from_rc_str(string: RcStr) -> InternedString {
 646         InternedString {
 647             string: string,
 648         }
 649     }
 650
 651     #[inline]
 652     pub fn new_from_name(name: ast::Name) -> InternedString {
 653         let interner = get_ident_interner();
 654         InternedString::new_from_rc_str(interner.get(name))
 655     }
 656 }
 657
 658 impl Deref for InternedString {
 659     type Target = str;
 660
 661     fn deref(&self) -> &str { &*self.string }
 662 }
 663
 664 impl fmt::Debug for InternedString {
 665     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
 666         fmt::Debug::fmt(&self.string, f)
 667     }
 668 }
 669
 670 impl fmt::Display for InternedString {
 671     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
 672         fmt::Display::fmt(&self.string, f)
 673     }
 674 }
 675
 676 impl<'a> PartialEq<&'a str> for InternedString {
 677     #[inline(always)]
 678     fn eq(&self, other: & &'a str) -> bool {
 679         PartialEq::eq(&self.string[..], *other)
 680     }
 681     #[inline(always)]
 682     fn ne(&self, other: & &'a str) -> bool {
 683         PartialEq::ne(&self.string[..], *other)
 684     }
 685 }
 686
 687 impl<'a> PartialEq<InternedString> for &'a str {
 688     #[inline(always)]
 689     fn eq(&self, other: &InternedString) -> bool {
 690         PartialEq::eq(*self, &other.string[..])
 691     }
 692     #[inline(always)]
 693     fn ne(&self, other: &InternedString) -> bool {
 694         PartialEq::ne(*self, &other.string[..])
 695     }
 696 }
 697
 698 impl Decodable for InternedString {
 699     fn decode<D: Decoder>(d: &mut D) -> Result<InternedString, D::Error> {
 700         Ok(intern(try!(d.read_str()).as_ref()).as_str())
 701     }
 702 }
 703
 704 impl Encodable for InternedString {
 705     fn encode<S: Encoder>(&self, s: &mut S) -> Result<(), S::Error> {
 706         s.emit_str(&self.string)
 707     }
 708 }
 709
 710 /// Interns and returns the string contents of an identifier, using the
 711 /// thread-local interner.
 712 #[inline]
 713 pub fn intern_and_get_ident(s: &str) -> InternedString {
 714     intern(s).as_str()
 715 }
 716
 717 /// Maps a string to its interned representation.
 718 #[inline]
 719 pub fn intern(s: &str) -> ast::Name {
 720     get_ident_interner().intern(s)
 721 }
 722
 723 /// gensym's a new usize, using the current interner.
 724 #[inline]
 725 pub fn gensym(s: &str) -> ast::Name {
 726     get_ident_interner().gensym(s)
 727 }
 728
 729 /// Maps a string to an identifier with an empty syntax context.
 730 #[inline]
 731 pub fn str_to_ident(s: &str) -> ast::Ident {
 732     ast::Ident::with_empty_ctxt(intern(s))
 733 }
 734
 735 /// Maps a string to a gensym'ed identifier.
 736 #[inline]
 737 pub fn gensym_ident(s: &str) -> ast::Ident {
 738     ast::Ident::with_empty_ctxt(gensym(s))
 739 }
 740
 741 // create a fresh name that maps to the same string as the old one.
 742 // note that this guarantees that str_ptr_eq(ident_to_string(src),interner_get(fresh_name(src)));
 743 // that is, that the new name and the old one are connected to ptr_eq strings.
 744 pub fn fresh_name(src: ast::Ident) -> ast::Name {
 745     let interner = get_ident_interner();
 746     interner.gensym_copy(src.name)
 747     // following: debug version. Could work in final except that it's incompatible with
 748     // good error messages and uses of struct names in ambiguous could-be-binding
 749     // locations. Also definitely destroys the guarantee given above about ptr_eq.
 750     /*let num = rand::thread_rng().gen_uint_range(0,0xffff);
 751     gensym(format!("{}_{}",ident_to_string(src),num))*/
 752 }
 753
 754 // create a fresh mark.
 755 pub fn fresh_mark() -> ast::Mrk {
 756     gensym("mark").0
 757 }
 758
 759 #[cfg(test)]
 760 mod tests {
 761     use super::*;
 762     use ast;
 763     use ext::mtwt;
 764
 765     fn mark_ident(id : ast::Ident, m : ast::Mrk) -> ast::Ident {
 766         ast::Ident::new(id.name, mtwt::apply_mark(m, id.ctxt))
 767     }
 768
 769     #[test] fn mtwt_token_eq_test() {
 770         assert!(Gt.mtwt_eq(&Gt));
 771         let a = str_to_ident("bac");
 772         let a1 = mark_ident(a,92);
 773         assert!(Ident(a, ModName).mtwt_eq(&Ident(a1, Plain)));
 774     }
 775 }