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.
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.
12 use ast::{P, Ident, Name, Mrk};
15 use util::interner::{RcStr, StrInterner};
18 use serialize::{Decodable, Decoder, Encodable, Encoder};
23 use std::path::BytesContainer;
25 #[allow(non_camel_case_types)]
26 #[deriving(Clone, Encodable, Decodable, Eq, IterBytes)]
40 #[allow(non_camel_case_types)]
41 #[deriving(Clone, Encodable, Decodable, Eq, IterBytes)]
43 /* Expression-operator symbols. */
58 /* Structural symbols */
82 LIT_INT(i64, ast::IntTy),
83 LIT_UINT(u64, ast::UintTy),
84 LIT_INT_UNSUFFIXED(i64),
85 LIT_FLOAT(ast::Ident, ast::FloatTy),
86 LIT_FLOAT_UNSUFFIXED(ast::Ident),
88 LIT_STR_RAW(ast::Ident, uint), /* raw str delimited by n hash symbols */
91 // an identifier contains an "is_mod_name" boolean,
92 // indicating whether :: follows this token with no
93 // whitespace in between.
94 IDENT(ast::Ident, bool),
98 /* For interpolation */
99 INTERPOLATED(Nonterminal),
101 DOC_COMMENT(ast::Ident),
105 #[deriving(Clone, Encodable, Decodable, Eq, IterBytes)]
106 /// For interpolation during macro expansion.
107 pub enum Nonterminal {
109 NtBlock(P<ast::Block>),
114 NtIdent(~ast::Ident, bool),
115 NtAttr(@ast::Attribute), // #[foo]
117 NtTT( @ast::TokenTree), // needs @ed to break a circularity
118 NtMatchers(~[ast::Matcher])
121 pub fn binop_to_str(o: BinOp) -> ~str {
136 pub fn to_str(t: &Token) -> ~str {
149 BINOP(op) => binop_to_str(op),
150 BINOPEQ(op) => binop_to_str(op) + "=",
152 /* Structural symbols */
177 char::from_u32(c).unwrap().escape_default(|c| {
184 i.to_str() + ast_util::int_ty_to_str(t)
187 u.to_str() + ast_util::uint_ty_to_str(t)
189 LIT_INT_UNSUFFIXED(i) => { i.to_str() }
191 let mut body = get_ident(s).get().to_str();
192 if body.ends_with(".") {
193 body.push_char('0'); // `10.f` is not a float literal
195 body + ast_util::float_ty_to_str(t)
197 LIT_FLOAT_UNSUFFIXED(s) => {
198 let mut body = get_ident(s).get().to_str();
199 if body.ends_with(".") {
200 body.push_char('0'); // `10.f` is not a float literal
205 format!("\"{}\"", get_ident(s).get().escape_default())
207 LIT_STR_RAW(s, n) => {
208 format!("r{delim}\"{string}\"{delim}",
209 delim="#".repeat(n), string=get_ident(s))
212 /* Name components */
213 IDENT(s, _) => get_ident(s).get().to_str(),
215 format!("'{}", get_ident(s))
220 DOC_COMMENT(s) => get_ident(s).get().to_str(),
222 INTERPOLATED(ref nt) => {
224 &NtExpr(e) => ::print::pprust::expr_to_str(e),
225 &NtAttr(e) => ::print::pprust::attribute_to_str(e),
227 ~"an interpolated " +
229 NtItem(..) => ~"item",
230 NtBlock(..) => ~"block",
231 NtStmt(..) => ~"statement",
232 NtPat(..) => ~"pattern",
233 NtAttr(..) => fail!("should have been handled"),
234 NtExpr(..) => fail!("should have been handled above"),
236 NtIdent(..) => ~"identifier",
237 NtPath(..) => ~"path",
239 NtMatchers(..) => ~"matcher sequence"
247 pub fn can_begin_expr(t: &Token) -> bool {
256 LIT_INT(_, _) => true,
257 LIT_UINT(_, _) => true,
258 LIT_INT_UNSUFFIXED(_) => true,
259 LIT_FLOAT(_, _) => true,
260 LIT_FLOAT_UNSUFFIXED(_) => true,
262 LIT_STR_RAW(_, _) => true,
266 BINOP(MINUS) => true,
269 BINOP(OR) => true, // in lambda syntax
270 OROR => true, // in lambda syntax
272 INTERPOLATED(NtExpr(..))
273 | INTERPOLATED(NtIdent(..))
274 | INTERPOLATED(NtBlock(..))
275 | INTERPOLATED(NtPath(..)) => true,
280 /// what's the opposite delimiter?
281 pub fn flip_delimiter(t: &token::Token) -> token::Token {
285 LBRACKET => RBRACKET,
288 RBRACKET => LBRACKET,
295 pub fn is_lit(t: &Token) -> bool {
298 LIT_INT(_, _) => true,
299 LIT_UINT(_, _) => true,
300 LIT_INT_UNSUFFIXED(_) => true,
301 LIT_FLOAT(_, _) => true,
302 LIT_FLOAT_UNSUFFIXED(_) => true,
304 LIT_STR_RAW(_, _) => true,
309 pub fn is_ident(t: &Token) -> bool {
310 match *t { IDENT(_, _) => true, _ => false }
313 pub fn is_ident_or_path(t: &Token) -> bool {
315 IDENT(_, _) | INTERPOLATED(NtPath(..)) => true,
320 pub fn is_plain_ident(t: &Token) -> bool {
321 match *t { IDENT(_, false) => true, _ => false }
324 pub fn is_bar(t: &Token) -> bool {
325 match *t { BINOP(OR) | OROR => true, _ => false }
328 // Get the first "argument"
330 ( $first:expr, $( $remainder:expr, )* ) => ( $first )
333 // Get the last "argument" (has to be done recursively to avoid phoney local ambiguity error)
335 ( $first:expr, $( $remainder:expr, )+ ) => ( last!( $( $remainder, )+ ) );
336 ( $first:expr, ) => ( $first )
339 // In this macro, there is the requirement that the name (the number) must be monotonically
340 // increasing by one in the special identifiers, starting at 0; the same holds for the keywords,
341 // except starting from the next number instead of zero, and with the additional exception that
342 // special identifiers are *also* allowed (they are deduplicated in the important place, the
343 // interner), an exception which is demonstrated by "static" and "self".
344 macro_rules! declare_special_idents_and_keywords {(
345 // So now, in these rules, why is each definition parenthesised?
346 // Answer: otherwise we get a spurious local ambiguity bug on the "}"
347 pub mod special_idents {
348 $( ($si_name:expr, $si_static:ident, $si_str:expr); )*
353 $( ($sk_name:expr, $sk_variant:ident, $sk_str:expr); )*
355 $( ($rk_name:expr, $rk_variant:ident, $rk_str:expr); )*
358 static STRICT_KEYWORD_START: Name = first!($( $sk_name, )*);
359 static STRICT_KEYWORD_FINAL: Name = last!($( $sk_name, )*);
360 static RESERVED_KEYWORD_START: Name = first!($( $rk_name, )*);
361 static RESERVED_KEYWORD_FINAL: Name = last!($( $rk_name, )*);
363 pub mod special_idents {
365 $( pub static $si_static: Ident = Ident { name: $si_name, ctxt: 0 }; )*
369 * All the valid words that have meaning in the Rust language.
371 * Rust keywords are either 'strict' or 'reserved'. Strict keywords may not
372 * appear as identifiers at all. Reserved keywords are not used anywhere in
373 * the language and may not appear as identifiers.
384 pub fn to_ident(&self) -> Ident {
386 $( $sk_variant => Ident { name: $sk_name, ctxt: 0 }, )*
387 $( $rk_variant => Ident { name: $rk_name, ctxt: 0 }, )*
393 fn mk_fresh_ident_interner() -> IdentInterner {
394 // The indices here must correspond to the numbers in
395 // special_idents, in Keyword to_ident(), and in static
403 interner::StrInterner::prefill(init_vec)
407 // If the special idents get renumbered, remember to modify these two as appropriate
408 static SELF_KEYWORD_NAME: Name = 1;
409 static STATIC_KEYWORD_NAME: Name = 2;
411 declare_special_idents_and_keywords! {
412 pub mod special_idents {
413 // These ones are statics
415 (super::SELF_KEYWORD_NAME, self_, "self");
416 (super::STATIC_KEYWORD_NAME, statik, "static");
420 (4, matchers, "matchers");
422 // outside of libsyntax
423 (5, clownshoe_abi, "__rust_abi");
424 (6, opaque, "<opaque>");
425 (7, unnamed_field, "<unnamed_field>");
426 (8, type_self, "Self");
430 // These ones are variants of the Keyword enum
434 (10, Break, "break");
435 (11, Const, "const");
436 (12, Crate, "crate");
439 (15, Extern, "extern");
440 (16, False, "false");
447 (23, __LogLevel, "__log_level");
449 (25, Match, "match");
456 (32, Return, "return");
457 // Static and Self are also special idents (prefill de-dupes)
458 (super::STATIC_KEYWORD_NAME, Static, "static");
459 (super::SELF_KEYWORD_NAME, Self, "self");
460 (33, Struct, "struct");
461 (34, Super, "super");
463 (36, Trait, "trait");
465 (38, Unsafe, "unsafe");
467 (40, While, "while");
468 (41, Continue, "continue");
473 (44, Alignof, "alignof");
475 (46, Offsetof, "offsetof");
477 (48, Sizeof, "sizeof");
478 (49, Typeof, "typeof");
479 (50, Unsized, "unsized");
480 (51, Yield, "yield");
486 * Maps a token to a record specifying the corresponding binary
489 pub fn token_to_binop(tok: &Token) -> Option<ast::BinOp> {
491 BINOP(STAR) => Some(ast::BiMul),
492 BINOP(SLASH) => Some(ast::BiDiv),
493 BINOP(PERCENT) => Some(ast::BiRem),
494 BINOP(PLUS) => Some(ast::BiAdd),
495 BINOP(MINUS) => Some(ast::BiSub),
496 BINOP(SHL) => Some(ast::BiShl),
497 BINOP(SHR) => Some(ast::BiShr),
498 BINOP(AND) => Some(ast::BiBitAnd),
499 BINOP(CARET) => Some(ast::BiBitXor),
500 BINOP(OR) => Some(ast::BiBitOr),
501 LT => Some(ast::BiLt),
502 LE => Some(ast::BiLe),
503 GE => Some(ast::BiGe),
504 GT => Some(ast::BiGt),
505 EQEQ => Some(ast::BiEq),
506 NE => Some(ast::BiNe),
507 ANDAND => Some(ast::BiAnd),
508 OROR => Some(ast::BiOr),
513 // looks like we can get rid of this completely...
514 pub type IdentInterner = StrInterner;
516 // if an interner exists in TLS, return it. Otherwise, prepare a
518 pub fn get_ident_interner() -> @IdentInterner {
519 local_data_key!(key: @::parse::token::IdentInterner)
520 match local_data::get(key, |k| k.map(|k| *k)) {
521 Some(interner) => interner,
523 let interner = @mk_fresh_ident_interner();
524 local_data::set(key, interner);
530 /// Represents a string stored in the task-local interner. Because the
531 /// interner lives for the life of the task, this can be safely treated as an
532 /// immortal string, as long as it never crosses between tasks.
534 /// FIXME(pcwalton): You must be careful about what you do in the destructors
535 /// of objects stored in TLS, because they may run after the interner is
536 /// destroyed. In particular, they must not access string contents. This can
537 /// be fixed in the future by just leaking all strings until task death
539 #[deriving(Clone, Eq, IterBytes, Ord, TotalEq, TotalOrd)]
540 pub struct InternedString {
544 impl InternedString {
546 pub fn new(string: &'static str) -> InternedString {
548 string: RcStr::new(string),
553 fn new_from_rc_str(string: RcStr) -> InternedString {
560 pub fn get<'a>(&'a self) -> &'a str {
561 self.string.as_slice()
565 impl BytesContainer for InternedString {
566 fn container_as_bytes<'a>(&'a self) -> &'a [u8] {
567 // FIXME(pcwalton): This is a workaround for the incorrect signature
568 // of `BytesContainer`, which is itself a workaround for the lack of
571 let this = self.get();
572 cast::transmute(this.container_as_bytes())
577 impl fmt::Show for InternedString {
578 fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
579 write!(f.buf, "{}", self.string.as_slice())
583 impl<'a> Equiv<&'a str> for InternedString {
584 fn equiv(&self, other: & &'a str) -> bool {
585 (*other) == self.string.as_slice()
589 impl<D:Decoder> Decodable<D> for InternedString {
590 fn decode(d: &mut D) -> InternedString {
591 get_name(get_ident_interner().intern(d.read_str()))
595 impl<E:Encoder> Encodable<E> for InternedString {
596 fn encode(&self, e: &mut E) {
597 e.emit_str(self.string.as_slice())
601 /// Returns the string contents of a name, using the task-local interner.
603 pub fn get_name(name: Name) -> InternedString {
604 let interner = get_ident_interner();
605 InternedString::new_from_rc_str(interner.get(name))
608 /// Returns the string contents of an identifier, using the task-local
611 pub fn get_ident(ident: Ident) -> InternedString {
615 /// Interns and returns the string contents of an identifier, using the
616 /// task-local interner.
618 pub fn intern_and_get_ident(s: &str) -> InternedString {
622 /// Maps a string to its interned representation.
624 pub fn intern(s: &str) -> Name {
625 get_ident_interner().intern(s)
628 /// gensym's a new uint, using the current interner.
630 pub fn gensym(s: &str) -> Name {
631 get_ident_interner().gensym(s)
634 /// Maps a string to an identifier with an empty syntax context.
636 pub fn str_to_ident(s: &str) -> ast::Ident {
637 ast::Ident::new(intern(s))
640 /// Maps a string to a gensym'ed identifier.
642 pub fn gensym_ident(s: &str) -> ast::Ident {
643 ast::Ident::new(gensym(s))
646 // create a fresh name that maps to the same string as the old one.
647 // note that this guarantees that str_ptr_eq(ident_to_str(src),interner_get(fresh_name(src)));
648 // that is, that the new name and the old one are connected to ptr_eq strings.
649 pub fn fresh_name(src: &ast::Ident) -> Name {
650 let interner = get_ident_interner();
651 interner.gensym_copy(src.name)
652 // following: debug version. Could work in final except that it's incompatible with
653 // good error messages and uses of struct names in ambiguous could-be-binding
654 // locations. Also definitely destroys the guarantee given above about ptr_eq.
655 /*let num = rand::rng().gen_uint_range(0,0xffff);
656 gensym(format!("{}_{}",ident_to_str(src),num))*/
659 // create a fresh mark.
660 pub fn fresh_mark() -> Mrk {
664 // See the macro above about the types of keywords
666 pub fn is_keyword(kw: keywords::Keyword, tok: &Token) -> bool {
668 token::IDENT(sid, false) => { kw.to_ident().name == sid.name }
673 pub fn is_any_keyword(tok: &Token) -> bool {
675 token::IDENT(sid, false) => match sid.name {
676 SELF_KEYWORD_NAME | STATIC_KEYWORD_NAME |
677 STRICT_KEYWORD_START .. RESERVED_KEYWORD_FINAL => true,
684 pub fn is_strict_keyword(tok: &Token) -> bool {
686 token::IDENT(sid, false) => match sid.name {
687 SELF_KEYWORD_NAME | STATIC_KEYWORD_NAME |
688 STRICT_KEYWORD_START .. STRICT_KEYWORD_FINAL => true,
695 pub fn is_reserved_keyword(tok: &Token) -> bool {
697 token::IDENT(sid, false) => match sid.name {
698 RESERVED_KEYWORD_START .. RESERVED_KEYWORD_FINAL => true,
705 pub fn mtwt_token_eq(t1 : &Token, t2 : &Token) -> bool {
707 (&IDENT(id1,_),&IDENT(id2,_)) =>
708 ast_util::mtwt_resolve(id1) == ast_util::mtwt_resolve(id2),
720 fn mark_ident(id : ast::Ident, m : ast::Mrk) -> ast::Ident {
721 ast::Ident{name:id.name,ctxt:ast_util::new_mark(m,id.ctxt)}
724 #[test] fn mtwt_token_eq_test() {
725 assert!(mtwt_token_eq(>,>));
726 let a = str_to_ident("bac");
727 let a1 = mark_ident(a,92);
728 assert!(mtwt_token_eq(&IDENT(a,true),&IDENT(a1,false)));