Alex Crichton | 44bffbc | 2017-05-19 17:51:59 -0700 | [diff] [blame] | 1 | extern crate unicode_xid; |
| 2 | |
Alex Crichton | 76a5cc8 | 2017-05-23 07:01:44 -0700 | [diff] [blame] | 3 | use std::ascii; |
Alex Crichton | 44bffbc | 2017-05-19 17:51:59 -0700 | [diff] [blame] | 4 | use std::borrow::Borrow; |
| 5 | use std::cell::RefCell; |
| 6 | use std::collections::HashMap; |
| 7 | use std::fmt; |
| 8 | use std::iter; |
| 9 | use std::ops; |
| 10 | use std::rc::Rc; |
| 11 | use std::str::FromStr; |
| 12 | use std::vec; |
| 13 | |
| 14 | use proc_macro; |
| 15 | use self::unicode_xid::UnicodeXID; |
| 16 | use synom::space::{skip_whitespace, block_comment, whitespace}; |
| 17 | use synom::{IResult}; |
| 18 | |
| 19 | use {TokenTree, TokenKind, Delimiter, OpKind}; |
| 20 | |
| 21 | #[derive(Clone)] |
| 22 | pub struct TokenStream { |
| 23 | inner: Vec<TokenTree>, |
| 24 | } |
| 25 | |
| 26 | #[derive(Debug)] |
| 27 | pub struct LexError; |
| 28 | |
| 29 | impl TokenStream { |
| 30 | pub fn empty() -> TokenStream { |
| 31 | TokenStream { inner: Vec::new() } |
| 32 | } |
| 33 | |
| 34 | pub fn is_empty(&self) -> bool { |
| 35 | self.inner.len() == 0 |
| 36 | } |
| 37 | } |
| 38 | |
| 39 | impl FromStr for TokenStream { |
| 40 | type Err = LexError; |
| 41 | |
| 42 | fn from_str(src: &str) -> Result<TokenStream, LexError> { |
| 43 | match token_stream(src) { |
| 44 | IResult::Done(input, output) => { |
| 45 | if skip_whitespace(input).len() != 0 { |
| 46 | Err(LexError) |
| 47 | } else { |
| 48 | Ok(output) |
| 49 | } |
| 50 | } |
| 51 | IResult::Error => Err(LexError), |
| 52 | } |
| 53 | } |
| 54 | } |
| 55 | |
| 56 | impl fmt::Display for TokenStream { |
| 57 | fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { |
| 58 | let mut joint = false; |
| 59 | for (i, tt) in self.inner.iter().enumerate() { |
| 60 | if i != 0 && !joint { |
| 61 | write!(f, " ")?; |
| 62 | } |
| 63 | joint = false; |
| 64 | match tt.kind { |
| 65 | TokenKind::Sequence(delim, ref stream) => { |
| 66 | let (start, end) = match delim { |
| 67 | Delimiter::Parenthesis => ("(", ")"), |
| 68 | Delimiter::Brace => ("{", "}"), |
| 69 | Delimiter::Bracket => ("[", "]"), |
| 70 | Delimiter::None => ("", ""), |
| 71 | }; |
Alex Crichton | 852d53d | 2017-05-19 19:25:08 -0700 | [diff] [blame] | 72 | if stream.0.inner.len() == 0 { |
| 73 | write!(f, "{} {}", start, end)? |
| 74 | } else { |
| 75 | write!(f, "{} {} {}", start, stream, end)? |
| 76 | } |
Alex Crichton | 44bffbc | 2017-05-19 17:51:59 -0700 | [diff] [blame] | 77 | } |
Alex Crichton | 9c2fb0a | 2017-05-26 08:49:31 -0700 | [diff] [blame] | 78 | TokenKind::Word(ref sym) => write!(f, "{}", sym.as_str())?, |
Alex Crichton | 44bffbc | 2017-05-19 17:51:59 -0700 | [diff] [blame] | 79 | TokenKind::Op(ch, ref op) => { |
| 80 | write!(f, "{}", ch)?; |
| 81 | match *op { |
| 82 | OpKind::Alone => {} |
| 83 | OpKind::Joint => joint = true, |
| 84 | } |
| 85 | } |
| 86 | TokenKind::Literal(ref literal) => { |
| 87 | write!(f, "{}", literal)?; |
| 88 | // handle comments |
| 89 | if (literal.0).0.starts_with("/") { |
| 90 | write!(f, "\n")?; |
| 91 | } |
| 92 | } |
| 93 | } |
| 94 | } |
| 95 | |
| 96 | Ok(()) |
| 97 | } |
| 98 | } |
| 99 | |
| 100 | impl From<proc_macro::TokenStream> for TokenStream { |
| 101 | fn from(inner: proc_macro::TokenStream) -> TokenStream { |
| 102 | inner.to_string().parse().expect("compiler token stream parse failed") |
| 103 | } |
| 104 | } |
| 105 | |
| 106 | impl From<TokenStream> for proc_macro::TokenStream { |
| 107 | fn from(inner: TokenStream) -> proc_macro::TokenStream { |
| 108 | inner.to_string().parse().expect("failed to parse to compiler tokens") |
| 109 | } |
| 110 | } |
| 111 | |
| 112 | |
| 113 | impl From<TokenTree> for TokenStream { |
| 114 | fn from(tree: TokenTree) -> TokenStream { |
| 115 | TokenStream { inner: vec![tree] } |
| 116 | } |
| 117 | } |
| 118 | |
| 119 | impl iter::FromIterator<TokenStream> for TokenStream { |
| 120 | fn from_iter<I: IntoIterator<Item=TokenStream>>(streams: I) -> Self { |
| 121 | let mut v = Vec::new(); |
| 122 | |
| 123 | for stream in streams.into_iter() { |
| 124 | v.extend(stream.inner); |
| 125 | } |
| 126 | |
| 127 | TokenStream { inner: v } |
| 128 | } |
| 129 | } |
| 130 | |
| 131 | pub type TokenIter = vec::IntoIter<TokenTree>; |
| 132 | |
| 133 | impl IntoIterator for TokenStream { |
| 134 | type Item = TokenTree; |
| 135 | type IntoIter = TokenIter; |
| 136 | |
| 137 | fn into_iter(self) -> TokenIter { |
| 138 | self.inner.into_iter() |
| 139 | } |
| 140 | } |
| 141 | |
| 142 | #[derive(Clone, Copy, Default)] |
| 143 | pub struct Span; |
| 144 | |
| 145 | impl Span { |
| 146 | pub fn call_site() -> Span { |
| 147 | Span |
| 148 | } |
| 149 | } |
| 150 | |
| 151 | #[derive(Copy, Clone)] |
| 152 | pub struct Symbol(usize); |
| 153 | |
| 154 | thread_local!(static SYMBOLS: RefCell<Interner> = RefCell::new(Interner::new())); |
| 155 | |
| 156 | impl<'a> From<&'a str> for Symbol { |
| 157 | fn from(string: &'a str) -> Symbol { |
| 158 | Symbol(SYMBOLS.with(|s| s.borrow_mut().intern(string))) |
| 159 | } |
| 160 | } |
| 161 | |
| 162 | impl ops::Deref for Symbol { |
| 163 | type Target = str; |
| 164 | |
| 165 | fn deref(&self) -> &str { |
| 166 | SYMBOLS.with(|interner| { |
| 167 | let interner = interner.borrow(); |
| 168 | let s = interner.get(self.0); |
| 169 | unsafe { |
| 170 | &*(s as *const str) |
| 171 | } |
| 172 | }) |
| 173 | } |
| 174 | } |
| 175 | |
| 176 | struct Interner { |
| 177 | string_to_index: HashMap<MyRc, usize>, |
| 178 | index_to_string: Vec<Rc<String>>, |
| 179 | } |
| 180 | |
| 181 | #[derive(Hash, Eq, PartialEq)] |
| 182 | struct MyRc(Rc<String>); |
| 183 | |
| 184 | impl Borrow<str> for MyRc { |
| 185 | fn borrow(&self) -> &str { |
| 186 | &self.0 |
| 187 | } |
| 188 | } |
| 189 | |
| 190 | impl Interner { |
| 191 | fn new() -> Interner { |
| 192 | Interner { |
| 193 | string_to_index: HashMap::new(), |
| 194 | index_to_string: Vec::new(), |
| 195 | } |
| 196 | } |
| 197 | |
| 198 | fn intern(&mut self, s: &str) -> usize { |
| 199 | if let Some(&idx) = self.string_to_index.get(s) { |
| 200 | return idx |
| 201 | } |
| 202 | let s = Rc::new(s.to_string()); |
| 203 | self.index_to_string.push(s.clone()); |
| 204 | self.string_to_index.insert(MyRc(s), self.index_to_string.len() - 1); |
| 205 | self.index_to_string.len() - 1 |
| 206 | } |
| 207 | |
| 208 | fn get(&self, idx: usize) -> &str { |
| 209 | &self.index_to_string[idx] |
| 210 | } |
| 211 | } |
| 212 | |
| 213 | #[derive(Clone)] |
| 214 | pub struct Literal(String); |
| 215 | |
Alex Crichton | 852d53d | 2017-05-19 19:25:08 -0700 | [diff] [blame] | 216 | impl Literal { |
Alex Crichton | 9c2fb0a | 2017-05-26 08:49:31 -0700 | [diff] [blame] | 217 | pub fn byte_char(byte: u8) -> Literal { |
Alex Crichton | 76a5cc8 | 2017-05-23 07:01:44 -0700 | [diff] [blame] | 218 | match byte { |
| 219 | 0 => Literal(format!("b'\\0'")), |
| 220 | b'\"' => Literal(format!("b'\"'")), |
| 221 | n => { |
| 222 | let mut escaped = "b'".to_string(); |
| 223 | escaped.extend(ascii::escape_default(n).map(|c| c as char)); |
| 224 | escaped.push('\''); |
| 225 | Literal(escaped) |
| 226 | } |
| 227 | } |
| 228 | } |
| 229 | |
Alex Crichton | 9c2fb0a | 2017-05-26 08:49:31 -0700 | [diff] [blame] | 230 | pub fn byte_string(bytes: &[u8]) -> Literal { |
Alex Crichton | 852d53d | 2017-05-19 19:25:08 -0700 | [diff] [blame] | 231 | let mut escaped = "b\"".to_string(); |
| 232 | for b in bytes { |
| 233 | match *b { |
| 234 | b'\0' => escaped.push_str(r"\0"), |
| 235 | b'\t' => escaped.push_str(r"\t"), |
| 236 | b'\n' => escaped.push_str(r"\n"), |
| 237 | b'\r' => escaped.push_str(r"\r"), |
| 238 | b'"' => escaped.push_str("\\\""), |
| 239 | b'\\' => escaped.push_str("\\\\"), |
| 240 | b'\x20' ... b'\x7E' => escaped.push(*b as char), |
| 241 | _ => escaped.push_str(&format!("\\x{:02X}", b)), |
| 242 | } |
| 243 | } |
| 244 | escaped.push('"'); |
| 245 | Literal(escaped) |
| 246 | } |
Alex Crichton | 76a5cc8 | 2017-05-23 07:01:44 -0700 | [diff] [blame] | 247 | |
| 248 | pub fn doccomment(s: &str) -> Literal { |
| 249 | Literal(s.to_string()) |
| 250 | } |
Alex Crichton | 9c2fb0a | 2017-05-26 08:49:31 -0700 | [diff] [blame] | 251 | |
| 252 | pub fn float(s: &str) -> Literal { |
| 253 | Literal(s.to_string()) |
| 254 | } |
| 255 | |
| 256 | pub fn integer(s: &str) -> Literal { |
| 257 | Literal(s.to_string()) |
| 258 | } |
Alex Crichton | 3131662 | 2017-05-26 12:54:47 -0700 | [diff] [blame] | 259 | |
| 260 | pub fn raw_string(s: &str, pounds: usize) -> Literal { |
| 261 | let mut ret = format!("r"); |
| 262 | ret.extend((0..pounds).map(|_| "#")); |
| 263 | ret.push('"'); |
| 264 | ret.push_str(s); |
| 265 | ret.push('"'); |
| 266 | ret.extend((0..pounds).map(|_| "#")); |
| 267 | Literal(ret) |
| 268 | } |
| 269 | |
| 270 | pub fn raw_byte_string(s: &str, pounds: usize) -> Literal { |
Alex Crichton | 7ed6d28 | 2017-05-26 13:42:50 -0700 | [diff] [blame^] | 271 | let mut ret = format!("br"); |
Alex Crichton | 3131662 | 2017-05-26 12:54:47 -0700 | [diff] [blame] | 272 | ret.extend((0..pounds).map(|_| "#")); |
| 273 | ret.push('"'); |
| 274 | ret.push_str(s); |
| 275 | ret.push('"'); |
| 276 | ret.extend((0..pounds).map(|_| "#")); |
| 277 | Literal(ret) |
| 278 | } |
Alex Crichton | 852d53d | 2017-05-19 19:25:08 -0700 | [diff] [blame] | 279 | } |
| 280 | |
Alex Crichton | 44bffbc | 2017-05-19 17:51:59 -0700 | [diff] [blame] | 281 | impl fmt::Display for Literal { |
| 282 | fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { |
| 283 | self.0.fmt(f) |
| 284 | } |
| 285 | } |
| 286 | |
Alex Crichton | 9c2fb0a | 2017-05-26 08:49:31 -0700 | [diff] [blame] | 287 | macro_rules! ints { |
Alex Crichton | 44bffbc | 2017-05-19 17:51:59 -0700 | [diff] [blame] | 288 | ($($t:ty,)*) => {$( |
| 289 | impl From<$t> for Literal { |
| 290 | fn from(t: $t) -> Literal { |
Alex Crichton | 852d53d | 2017-05-19 19:25:08 -0700 | [diff] [blame] | 291 | Literal(format!(concat!("{}", stringify!($t)), t)) |
Alex Crichton | 44bffbc | 2017-05-19 17:51:59 -0700 | [diff] [blame] | 292 | } |
| 293 | } |
| 294 | )*} |
| 295 | } |
| 296 | |
Alex Crichton | 9c2fb0a | 2017-05-26 08:49:31 -0700 | [diff] [blame] | 297 | ints! { |
Alex Crichton | 852d53d | 2017-05-19 19:25:08 -0700 | [diff] [blame] | 298 | u8, u16, u32, u64, usize, |
| 299 | i8, i16, i32, i64, isize, |
Alex Crichton | 9c2fb0a | 2017-05-26 08:49:31 -0700 | [diff] [blame] | 300 | } |
| 301 | |
| 302 | macro_rules! floats { |
| 303 | ($($t:ty,)*) => {$( |
| 304 | impl From<$t> for Literal { |
| 305 | fn from(t: $t) -> Literal { |
| 306 | assert!(!t.is_nan()); |
| 307 | assert!(!t.is_infinite()); |
| 308 | Literal(format!(concat!("{}", stringify!($t)), t)) |
| 309 | } |
| 310 | } |
| 311 | )*} |
| 312 | } |
| 313 | |
| 314 | floats! { |
Alex Crichton | 852d53d | 2017-05-19 19:25:08 -0700 | [diff] [blame] | 315 | f32, f64, |
| 316 | } |
| 317 | |
Alex Crichton | 44bffbc | 2017-05-19 17:51:59 -0700 | [diff] [blame] | 318 | impl<'a> From<&'a str> for Literal { |
| 319 | fn from(t: &'a str) -> Literal { |
| 320 | let mut s = t.chars().flat_map(|c| c.escape_default()).collect::<String>(); |
| 321 | s.push('"'); |
| 322 | s.insert(0, '"'); |
| 323 | Literal(s) |
| 324 | } |
| 325 | } |
| 326 | |
| 327 | impl From<char> for Literal { |
| 328 | fn from(t: char) -> Literal { |
| 329 | Literal(format!("'{}'", t.escape_default())) |
| 330 | } |
| 331 | } |
| 332 | |
| 333 | named!(token_stream -> TokenStream, |
| 334 | map!(token_trees, |s: Vec<TokenTree>| TokenStream { inner: s })); |
| 335 | |
| 336 | named!(token_trees -> Vec<TokenTree>, many0!(token_tree)); |
| 337 | |
| 338 | named!(token_tree -> TokenTree, |
| 339 | map!(token_kind, |s: TokenKind| { |
| 340 | TokenTree { |
| 341 | span: ::Span(Span), |
| 342 | kind: s, |
| 343 | } |
| 344 | })); |
| 345 | |
| 346 | named!(token_kind -> TokenKind, alt!( |
| 347 | map!(delimited, |(d, s): (Delimiter, TokenStream)| { |
| 348 | TokenKind::Sequence(d, ::TokenStream(s)) |
| 349 | }) |
| 350 | | |
Michael Layzell | bad5801 | 2017-05-24 20:42:52 -0400 | [diff] [blame] | 351 | map!(literal, |l| TokenKind::Literal(::Literal(l))) // must be before symbol |
Alex Crichton | 44bffbc | 2017-05-19 17:51:59 -0700 | [diff] [blame] | 352 | | |
Michael Layzell | bad5801 | 2017-05-24 20:42:52 -0400 | [diff] [blame] | 353 | map!(symbol, |w| TokenKind::Word(::Symbol(w))) |
Alex Crichton | 44bffbc | 2017-05-19 17:51:59 -0700 | [diff] [blame] | 354 | | |
| 355 | map!(op, |(op, kind): (char, OpKind)| { |
| 356 | TokenKind::Op(op, kind) |
| 357 | }) |
| 358 | )); |
| 359 | |
| 360 | named!(delimited -> (Delimiter, TokenStream), alt!( |
| 361 | delimited!( |
| 362 | punct!("("), |
| 363 | token_stream, |
| 364 | punct!(")") |
| 365 | ) => { |ts| (Delimiter::Parenthesis, ts) } |
| 366 | | |
| 367 | delimited!( |
| 368 | punct!("["), |
| 369 | token_stream, |
| 370 | punct!("]") |
| 371 | ) => { |ts| (Delimiter::Bracket, ts) } |
| 372 | | |
| 373 | delimited!( |
| 374 | punct!("{"), |
| 375 | token_stream, |
| 376 | punct!("}") |
| 377 | ) => { |ts| (Delimiter::Brace, ts) } |
| 378 | )); |
| 379 | |
| 380 | named!(symbol -> Symbol, alt!( |
| 381 | lifetime |
| 382 | | |
| 383 | map!(word, Symbol::from) |
| 384 | )); |
| 385 | |
| 386 | named!(lifetime -> Symbol, preceded!( |
| 387 | punct!("'"), |
| 388 | alt!( |
| 389 | // TODO: can we get rid of this allocation? |
| 390 | map!(word, |id| Symbol::from(&format!("'{}", id)[..])) |
| 391 | | |
| 392 | map!(keyword!("static"), |_| Symbol::from("'static")) |
| 393 | ) |
| 394 | )); |
| 395 | |
| 396 | fn word(mut input: &str) -> IResult<&str, &str> { |
| 397 | input = skip_whitespace(input); |
| 398 | |
| 399 | let mut chars = input.char_indices(); |
| 400 | match chars.next() { |
| 401 | Some((_, ch)) if UnicodeXID::is_xid_start(ch) || ch == '_' => {} |
| 402 | _ => return IResult::Error, |
| 403 | } |
| 404 | |
| 405 | for (i, ch) in chars { |
| 406 | if !UnicodeXID::is_xid_continue(ch) { |
| 407 | return IResult::Done(&input[i..], &input[..i]) |
| 408 | } |
| 409 | } |
| 410 | |
| 411 | IResult::Done("", input) |
| 412 | } |
| 413 | |
| 414 | fn literal(input: &str) -> IResult<&str, Literal> { |
| 415 | let input_no_ws = skip_whitespace(input); |
| 416 | |
| 417 | match literal_nocapture(input_no_ws) { |
| 418 | IResult::Done(a, ()) => { |
| 419 | let start = input.len() - input_no_ws.len(); |
| 420 | let len = input_no_ws.len() - a.len(); |
| 421 | let end = start + len; |
Alex Crichton | 44bffbc | 2017-05-19 17:51:59 -0700 | [diff] [blame] | 422 | IResult::Done(a, Literal(input[start..end].to_string())) |
| 423 | } |
| 424 | IResult::Error => IResult::Error, |
| 425 | } |
| 426 | } |
| 427 | |
| 428 | named!(literal_nocapture -> (), alt!( |
| 429 | string |
| 430 | | |
| 431 | byte_string |
| 432 | | |
| 433 | byte |
| 434 | | |
| 435 | character |
| 436 | | |
| 437 | float |
| 438 | | |
| 439 | int |
| 440 | | |
| 441 | boolean |
| 442 | | |
| 443 | doc_comment |
| 444 | )); |
| 445 | |
| 446 | named!(string -> (), alt!( |
| 447 | quoted_string |
| 448 | | |
| 449 | preceded!( |
| 450 | punct!("r"), |
| 451 | raw_string |
| 452 | ) => { |_| () } |
| 453 | )); |
| 454 | |
| 455 | named!(quoted_string -> (), delimited!( |
| 456 | punct!("\""), |
| 457 | cooked_string, |
| 458 | tag!("\"") |
| 459 | )); |
| 460 | |
| 461 | fn cooked_string(input: &str) -> IResult<&str, ()> { |
| 462 | let mut chars = input.char_indices().peekable(); |
| 463 | while let Some((byte_offset, ch)) = chars.next() { |
| 464 | match ch { |
| 465 | '"' => { |
| 466 | return IResult::Done(&input[byte_offset..], ()); |
| 467 | } |
| 468 | '\r' => { |
| 469 | if let Some((_, '\n')) = chars.next() { |
| 470 | // ... |
| 471 | } else { |
| 472 | break; |
| 473 | } |
| 474 | } |
| 475 | '\\' => { |
| 476 | match chars.next() { |
| 477 | Some((_, 'x')) => { |
| 478 | if !backslash_x_char(&mut chars) { |
| 479 | break |
| 480 | } |
| 481 | } |
| 482 | Some((_, 'n')) | |
| 483 | Some((_, 'r')) | |
| 484 | Some((_, 't')) | |
| 485 | Some((_, '\\')) | |
| 486 | Some((_, '\'')) | |
| 487 | Some((_, '"')) | |
| 488 | Some((_, '0')) => {} |
| 489 | Some((_, 'u')) => { |
| 490 | if !backslash_u(&mut chars) { |
| 491 | break |
| 492 | } |
| 493 | } |
| 494 | Some((_, '\n')) | Some((_, '\r')) => { |
| 495 | while let Some(&(_, ch)) = chars.peek() { |
| 496 | if ch.is_whitespace() { |
| 497 | chars.next(); |
| 498 | } else { |
| 499 | break; |
| 500 | } |
| 501 | } |
| 502 | } |
| 503 | _ => break, |
| 504 | } |
| 505 | } |
| 506 | _ch => {} |
| 507 | } |
| 508 | } |
| 509 | IResult::Error |
| 510 | } |
| 511 | |
| 512 | named!(byte_string -> (), alt!( |
| 513 | delimited!( |
| 514 | punct!("b\""), |
| 515 | cooked_byte_string, |
| 516 | tag!("\"") |
| 517 | ) => { |_| () } |
| 518 | | |
| 519 | preceded!( |
| 520 | punct!("br"), |
| 521 | raw_string |
| 522 | ) => { |_| () } |
| 523 | )); |
| 524 | |
| 525 | fn cooked_byte_string(mut input: &str) -> IResult<&str, ()> { |
| 526 | let mut bytes = input.bytes().enumerate(); |
| 527 | 'outer: while let Some((offset, b)) = bytes.next() { |
| 528 | match b { |
| 529 | b'"' => { |
| 530 | return IResult::Done(&input[offset..], ()); |
| 531 | } |
| 532 | b'\r' => { |
| 533 | if let Some((_, b'\n')) = bytes.next() { |
| 534 | // ... |
| 535 | } else { |
| 536 | break; |
| 537 | } |
| 538 | } |
| 539 | b'\\' => { |
| 540 | match bytes.next() { |
| 541 | Some((_, b'x')) => { |
| 542 | if !backslash_x_byte(&mut bytes) { |
| 543 | break |
| 544 | } |
| 545 | } |
| 546 | Some((_, b'n')) | |
| 547 | Some((_, b'r')) | |
| 548 | Some((_, b't')) | |
| 549 | Some((_, b'\\')) | |
| 550 | Some((_, b'0')) | |
| 551 | Some((_, b'\'')) | |
| 552 | Some((_, b'"')) => {} |
| 553 | Some((newline, b'\n')) | |
| 554 | Some((newline, b'\r')) => { |
| 555 | let rest = &input[newline + 1..]; |
| 556 | for (offset, ch) in rest.char_indices() { |
| 557 | if !ch.is_whitespace() { |
| 558 | input = &rest[offset..]; |
| 559 | bytes = input.bytes().enumerate(); |
| 560 | continue 'outer; |
| 561 | } |
| 562 | } |
| 563 | break; |
| 564 | } |
| 565 | _ => break, |
| 566 | } |
| 567 | } |
| 568 | b if b < 0x80 => {} |
| 569 | _ => break, |
| 570 | } |
| 571 | } |
| 572 | IResult::Error |
| 573 | } |
| 574 | |
| 575 | fn raw_string(input: &str) -> IResult<&str, ()> { |
| 576 | let mut chars = input.char_indices(); |
| 577 | let mut n = 0; |
| 578 | while let Some((byte_offset, ch)) = chars.next() { |
| 579 | match ch { |
| 580 | '"' => { |
| 581 | n = byte_offset; |
| 582 | break; |
| 583 | } |
| 584 | '#' => {} |
| 585 | _ => return IResult::Error, |
| 586 | } |
| 587 | } |
| 588 | for (byte_offset, ch) in chars { |
| 589 | match ch { |
| 590 | '"' if input[byte_offset + 1..].starts_with(&input[..n]) => { |
| 591 | let rest = &input[byte_offset + 1 + n..]; |
| 592 | return IResult::Done(rest, ()) |
| 593 | } |
| 594 | '\r' => {} |
| 595 | _ => {} |
| 596 | } |
| 597 | } |
| 598 | IResult::Error |
| 599 | } |
| 600 | |
| 601 | named!(byte -> (), do_parse!( |
| 602 | punct!("b") >> |
| 603 | tag!("'") >> |
| 604 | cooked_byte >> |
| 605 | tag!("'") >> |
| 606 | (()) |
| 607 | )); |
| 608 | |
| 609 | fn cooked_byte(input: &str) -> IResult<&str, ()> { |
| 610 | let mut bytes = input.bytes().enumerate(); |
| 611 | let ok = match bytes.next().map(|(_, b)| b) { |
| 612 | Some(b'\\') => { |
| 613 | match bytes.next().map(|(_, b)| b) { |
| 614 | Some(b'x') => backslash_x_byte(&mut bytes), |
| 615 | Some(b'n') | |
| 616 | Some(b'r') | |
| 617 | Some(b't') | |
| 618 | Some(b'\\') | |
| 619 | Some(b'0') | |
| 620 | Some(b'\'') | |
| 621 | Some(b'"') => true, |
| 622 | _ => false, |
| 623 | } |
| 624 | } |
| 625 | b => b.is_some(), |
| 626 | }; |
| 627 | if ok { |
| 628 | match bytes.next() { |
| 629 | Some((offset, _)) => IResult::Done(&input[offset..], ()), |
| 630 | None => IResult::Done("", ()), |
| 631 | } |
| 632 | } else { |
| 633 | IResult::Error |
| 634 | } |
| 635 | } |
| 636 | |
| 637 | named!(character -> (), do_parse!( |
| 638 | punct!("'") >> |
| 639 | cooked_char >> |
| 640 | tag!("'") >> |
| 641 | (()) |
| 642 | )); |
| 643 | |
| 644 | fn cooked_char(input: &str) -> IResult<&str, ()> { |
| 645 | let mut chars = input.char_indices(); |
| 646 | let ok = match chars.next().map(|(_, ch)| ch) { |
| 647 | Some('\\') => { |
| 648 | match chars.next().map(|(_, ch)| ch) { |
| 649 | Some('x') => backslash_x_char(&mut chars), |
| 650 | Some('u') => backslash_u(&mut chars), |
| 651 | Some('n') | |
| 652 | Some('r') | |
| 653 | Some('t') | |
| 654 | Some('\\') | |
| 655 | Some('0') | |
| 656 | Some('\'') | |
| 657 | Some('"') => true, |
| 658 | _ => false, |
| 659 | } |
| 660 | } |
| 661 | ch => ch.is_some(), |
| 662 | }; |
| 663 | if ok { |
| 664 | IResult::Done(chars.as_str(), ()) |
| 665 | } else { |
| 666 | IResult::Error |
| 667 | } |
| 668 | } |
| 669 | |
| 670 | macro_rules! next_ch { |
| 671 | ($chars:ident @ $pat:pat $(| $rest:pat)*) => { |
| 672 | match $chars.next() { |
| 673 | Some((_, ch)) => match ch { |
| 674 | $pat $(| $rest)* => ch, |
| 675 | _ => return false, |
| 676 | }, |
| 677 | None => return false |
| 678 | } |
| 679 | }; |
| 680 | } |
| 681 | |
| 682 | fn backslash_x_char<I>(chars: &mut I) -> bool |
| 683 | where I: Iterator<Item = (usize, char)> |
| 684 | { |
| 685 | next_ch!(chars @ '0'...'7'); |
| 686 | next_ch!(chars @ '0'...'9' | 'a'...'f' | 'A'...'F'); |
| 687 | true |
| 688 | } |
| 689 | |
| 690 | fn backslash_x_byte<I>(chars: &mut I) -> bool |
| 691 | where I: Iterator<Item = (usize, u8)> |
| 692 | { |
| 693 | next_ch!(chars @ b'0'...b'9' | b'a'...b'f' | b'A'...b'F'); |
| 694 | next_ch!(chars @ b'0'...b'9' | b'a'...b'f' | b'A'...b'F'); |
| 695 | true |
| 696 | } |
| 697 | |
| 698 | fn backslash_u<I>(chars: &mut I) -> bool |
| 699 | where I: Iterator<Item = (usize, char)> |
| 700 | { |
| 701 | next_ch!(chars @ '{'); |
| 702 | next_ch!(chars @ '0'...'9' | 'a'...'f' | 'A'...'F'); |
| 703 | let b = next_ch!(chars @ '0'...'9' | 'a'...'f' | 'A'...'F' | '}'); |
| 704 | if b == '}' { |
| 705 | return true |
| 706 | } |
| 707 | let c = next_ch!(chars @ '0'...'9' | 'a'...'f' | 'A'...'F' | '}'); |
| 708 | if c == '}' { |
| 709 | return true |
| 710 | } |
| 711 | let d = next_ch!(chars @ '0'...'9' | 'a'...'f' | 'A'...'F' | '}'); |
| 712 | if d == '}' { |
| 713 | return true |
| 714 | } |
| 715 | let e = next_ch!(chars @ '0'...'9' | 'a'...'f' | 'A'...'F' | '}'); |
| 716 | if e == '}' { |
| 717 | return true |
| 718 | } |
| 719 | let f = next_ch!(chars @ '0'...'9' | 'a'...'f' | 'A'...'F' | '}'); |
| 720 | if f == '}' { |
| 721 | return true |
| 722 | } |
| 723 | next_ch!(chars @ '}'); |
| 724 | true |
| 725 | } |
| 726 | |
| 727 | named!(float -> (), do_parse!( |
| 728 | float_string >> |
| 729 | alt!( |
| 730 | tag!("f32") => { |_| () } |
| 731 | | |
| 732 | tag!("f64") => { |_| () } |
| 733 | | |
| 734 | epsilon!() |
| 735 | ) >> |
| 736 | (()) |
| 737 | )); |
| 738 | |
| 739 | fn float_string(input: &str) -> IResult<&str, ()> { |
| 740 | let mut chars = input.chars().peekable(); |
| 741 | match chars.next() { |
| 742 | Some(ch) if ch >= '0' && ch <= '9' => {} |
| 743 | _ => return IResult::Error, |
| 744 | } |
| 745 | |
| 746 | let mut len = 1; |
| 747 | let mut has_dot = false; |
| 748 | let mut has_exp = false; |
| 749 | while let Some(&ch) = chars.peek() { |
| 750 | match ch { |
| 751 | '0'...'9' | '_' => { |
| 752 | chars.next(); |
| 753 | len += 1; |
| 754 | } |
| 755 | '.' => { |
| 756 | if has_dot { |
| 757 | break; |
| 758 | } |
| 759 | chars.next(); |
| 760 | if chars.peek() |
| 761 | .map(|&ch| ch == '.' || UnicodeXID::is_xid_start(ch)) |
| 762 | .unwrap_or(false) { |
| 763 | return IResult::Error; |
| 764 | } |
| 765 | len += 1; |
| 766 | has_dot = true; |
| 767 | } |
| 768 | 'e' | 'E' => { |
| 769 | chars.next(); |
| 770 | len += 1; |
| 771 | has_exp = true; |
| 772 | break; |
| 773 | } |
| 774 | _ => break, |
| 775 | } |
| 776 | } |
| 777 | |
| 778 | let rest = &input[len..]; |
| 779 | if !(has_dot || has_exp || rest.starts_with("f32") || rest.starts_with("f64")) { |
| 780 | return IResult::Error; |
| 781 | } |
| 782 | |
| 783 | if has_exp { |
| 784 | let mut has_exp_value = false; |
| 785 | while let Some(&ch) = chars.peek() { |
| 786 | match ch { |
| 787 | '+' | '-' => { |
| 788 | if has_exp_value { |
| 789 | break; |
| 790 | } |
| 791 | chars.next(); |
| 792 | len += 1; |
| 793 | } |
| 794 | '0'...'9' => { |
| 795 | chars.next(); |
| 796 | len += 1; |
| 797 | has_exp_value = true; |
| 798 | } |
| 799 | '_' => { |
| 800 | chars.next(); |
| 801 | len += 1; |
| 802 | } |
| 803 | _ => break, |
| 804 | } |
| 805 | } |
| 806 | if !has_exp_value { |
| 807 | return IResult::Error; |
| 808 | } |
| 809 | } |
| 810 | |
| 811 | IResult::Done(&input[len..], ()) |
| 812 | } |
| 813 | |
| 814 | named!(int -> (), do_parse!( |
| 815 | digits >> |
| 816 | alt!( |
| 817 | tag!("isize") => { |_| () } |
| 818 | | |
| 819 | tag!("i8") => { |_| () } |
| 820 | | |
| 821 | tag!("i16") => { |_| () } |
| 822 | | |
| 823 | tag!("i32") => { |_| () } |
| 824 | | |
| 825 | tag!("i64") => { |_| () } |
| 826 | | |
| 827 | tag!("usize") => { |_| () } |
| 828 | | |
| 829 | tag!("u8") => { |_| () } |
| 830 | | |
| 831 | tag!("u16") => { |_| () } |
| 832 | | |
| 833 | tag!("u32") => { |_| () } |
| 834 | | |
| 835 | tag!("u64") => { |_| () } |
| 836 | | |
| 837 | epsilon!() |
| 838 | ) >> |
| 839 | (()) |
| 840 | )); |
| 841 | |
| 842 | fn digits(mut input: &str) -> IResult<&str, ()> { |
| 843 | let base = if input.starts_with("0x") { |
| 844 | input = &input[2..]; |
| 845 | 16 |
| 846 | } else if input.starts_with("0o") { |
| 847 | input = &input[2..]; |
| 848 | 8 |
| 849 | } else if input.starts_with("0b") { |
| 850 | input = &input[2..]; |
| 851 | 2 |
| 852 | } else { |
| 853 | 10 |
| 854 | }; |
| 855 | |
| 856 | let mut value = 0u64; |
| 857 | let mut len = 0; |
| 858 | let mut empty = true; |
| 859 | for b in input.bytes() { |
| 860 | let digit = match b { |
| 861 | b'0'...b'9' => (b - b'0') as u64, |
| 862 | b'a'...b'f' => 10 + (b - b'a') as u64, |
| 863 | b'A'...b'F' => 10 + (b - b'A') as u64, |
| 864 | b'_' => { |
| 865 | if empty && base == 10 { |
| 866 | return IResult::Error; |
| 867 | } |
| 868 | len += 1; |
| 869 | continue; |
| 870 | } |
| 871 | _ => break, |
| 872 | }; |
| 873 | if digit >= base { |
| 874 | return IResult::Error; |
| 875 | } |
| 876 | value = match value.checked_mul(base) { |
| 877 | Some(value) => value, |
| 878 | None => return IResult::Error, |
| 879 | }; |
| 880 | value = match value.checked_add(digit) { |
| 881 | Some(value) => value, |
| 882 | None => return IResult::Error, |
| 883 | }; |
| 884 | len += 1; |
| 885 | empty = false; |
| 886 | } |
| 887 | if empty { |
| 888 | IResult::Error |
| 889 | } else { |
| 890 | IResult::Done(&input[len..], ()) |
| 891 | } |
| 892 | } |
| 893 | |
| 894 | named!(boolean -> (), alt!( |
| 895 | keyword!("true") => { |_| () } |
| 896 | | |
| 897 | keyword!("false") => { |_| () } |
| 898 | )); |
| 899 | |
| 900 | macro_rules! punct1 { |
| 901 | ($i:expr, $punct:expr) => { |
| 902 | punct1($i, $punct) |
| 903 | } |
| 904 | } |
| 905 | |
| 906 | fn punct1<'a>(input: &'a str, token: &'static str) -> IResult<&'a str, char> { |
| 907 | let input = skip_whitespace(input); |
| 908 | if input.starts_with(token) { |
| 909 | IResult::Done(&input[1..], token.chars().next().unwrap()) |
| 910 | } else { |
| 911 | IResult::Error |
| 912 | } |
| 913 | } |
| 914 | |
| 915 | named!(op -> (char, OpKind), alt!( |
| 916 | keyword!("_") => { |_| ('_', OpKind::Alone) } |
| 917 | | |
| 918 | punct1!("&&") => { |_| ('&', OpKind::Joint) } |
| 919 | | |
| 920 | punct1!("||") => { |_| ('|', OpKind::Joint) } |
| 921 | | |
| 922 | punct1!("->") => { |_| ('-', OpKind::Joint) } |
| 923 | | |
| 924 | punct1!("<-") => { |_| ('<', OpKind::Joint) } |
| 925 | | |
| 926 | punct1!("=>") => { |_| ('=', OpKind::Joint) } |
| 927 | | |
| 928 | punct1!("...") => { |_| ('.', OpKind::Joint) } |
| 929 | | |
| 930 | punct1!("..") => { |_| ('.', OpKind::Joint) } |
| 931 | | |
| 932 | punct!(".") => { |_| ('.', OpKind::Alone) } |
| 933 | | |
Alex Crichton | 852d53d | 2017-05-19 19:25:08 -0700 | [diff] [blame] | 934 | punct!(",") => { |_| (',', OpKind::Alone) } |
| 935 | | |
Alex Crichton | 44bffbc | 2017-05-19 17:51:59 -0700 | [diff] [blame] | 936 | bin_op_eq |
| 937 | | |
| 938 | bin_op |
| 939 | | |
| 940 | punct1!("<=") => { |_| ('<', OpKind::Joint) } |
| 941 | | |
| 942 | punct1!("==") => { |_| ('=', OpKind::Joint) } |
| 943 | | |
| 944 | punct1!("!=") => { |_| ('!', OpKind::Joint) } |
| 945 | | |
| 946 | punct1!(">=") => { |_| ('>', OpKind::Joint) } |
| 947 | | |
| 948 | punct1!("::") => { |_| (':', OpKind::Joint) } |
| 949 | | |
| 950 | punct!("=") => { |_| ('=', OpKind::Alone) } |
| 951 | | |
| 952 | punct!("<") => { |_| ('<', OpKind::Alone) } |
| 953 | | |
| 954 | punct!(">") => { |_| ('>', OpKind::Alone) } |
| 955 | | |
| 956 | punct!("!") => { |_| ('!', OpKind::Alone) } |
| 957 | | |
| 958 | punct!("~") => { |_| ('~', OpKind::Alone) } |
| 959 | | |
| 960 | punct!("@") => { |_| ('@', OpKind::Alone) } |
| 961 | | |
| 962 | punct!(",") => { |_| (',', OpKind::Alone) } |
| 963 | | |
| 964 | punct!(";") => { |_| (';', OpKind::Alone) } |
| 965 | | |
| 966 | punct!(":") => { |_| (':', OpKind::Alone) } |
| 967 | | |
| 968 | punct!("#") => { |_| ('#', OpKind::Alone) } |
| 969 | | |
| 970 | punct!("$") => { |_| ('$', OpKind::Alone) } |
| 971 | | |
| 972 | punct!("?") => { |_| ('?', OpKind::Alone) } |
| 973 | )); |
| 974 | |
| 975 | named!(bin_op -> (char, OpKind), alt!( |
| 976 | punct!("+") => { |_| ('+', OpKind::Alone) } |
| 977 | | |
| 978 | punct!("-") => { |_| ('-', OpKind::Alone) } |
| 979 | | |
| 980 | punct!("*") => { |_| ('*', OpKind::Alone) } |
| 981 | | |
| 982 | punct!("/") => { |_| ('/', OpKind::Alone) } |
| 983 | | |
| 984 | punct!("%") => { |_| ('%', OpKind::Alone) } |
| 985 | | |
| 986 | punct!("^") => { |_| ('^', OpKind::Alone) } |
| 987 | | |
| 988 | punct!("&") => { |_| ('&', OpKind::Alone) } |
| 989 | | |
| 990 | punct!("|") => { |_| ('|', OpKind::Alone) } |
| 991 | | |
| 992 | punct1!("<<") => { |_| ('<', OpKind::Joint) } |
| 993 | | |
| 994 | punct1!(">>") => { |_| ('>', OpKind::Joint) } |
| 995 | )); |
| 996 | |
| 997 | named!(bin_op_eq -> (char, OpKind), alt!( |
| 998 | punct1!("+=") => { |_| ('+', OpKind::Joint) } |
| 999 | | |
| 1000 | punct1!("-=") => { |_| ('-', OpKind::Joint) } |
| 1001 | | |
| 1002 | punct1!("*=") => { |_| ('*', OpKind::Joint) } |
| 1003 | | |
| 1004 | punct1!("/=") => { |_| ('/', OpKind::Joint) } |
| 1005 | | |
| 1006 | punct1!("%=") => { |_| ('%', OpKind::Joint) } |
| 1007 | | |
| 1008 | punct1!("^=") => { |_| ('^', OpKind::Joint) } |
| 1009 | | |
| 1010 | punct1!("&=") => { |_| ('&', OpKind::Joint) } |
| 1011 | | |
| 1012 | punct1!("|=") => { |_| ('|', OpKind::Joint) } |
| 1013 | | |
| 1014 | punct1!("<<=") => { |_| ('<', OpKind::Joint) } |
| 1015 | | |
| 1016 | punct1!(">>=") => { |_| ('>', OpKind::Joint) } |
| 1017 | )); |
| 1018 | |
| 1019 | named!(doc_comment -> (), alt!( |
| 1020 | do_parse!( |
| 1021 | punct!("//!") >> |
| 1022 | take_until!("\n") >> |
| 1023 | (()) |
| 1024 | ) |
| 1025 | | |
| 1026 | do_parse!( |
| 1027 | option!(whitespace) >> |
| 1028 | peek!(tag!("/*!")) >> |
| 1029 | block_comment >> |
| 1030 | (()) |
| 1031 | ) |
| 1032 | | |
| 1033 | do_parse!( |
| 1034 | punct!("///") >> |
| 1035 | not!(tag!("/")) >> |
| 1036 | take_until!("\n") >> |
| 1037 | (()) |
| 1038 | ) |
| 1039 | | |
| 1040 | do_parse!( |
| 1041 | option!(whitespace) >> |
| 1042 | peek!(tuple!(tag!("/**"), not!(tag!("*")))) >> |
| 1043 | block_comment >> |
| 1044 | (()) |
| 1045 | ) |
| 1046 | )); |
| 1047 | |
| 1048 | #[cfg(test)] |
| 1049 | mod tests { |
| 1050 | use super::*; |
| 1051 | |
| 1052 | #[test] |
| 1053 | fn symbols() { |
| 1054 | assert_eq!(&*Symbol::from("foo"), "foo"); |
| 1055 | assert_eq!(&*Symbol::from("bar"), "bar"); |
| 1056 | } |
| 1057 | |
| 1058 | #[test] |
| 1059 | fn literals() { |
| 1060 | assert_eq!(Literal::from("foo").to_string(), "\"foo\""); |
| 1061 | assert_eq!(Literal::from("\"").to_string(), "\"\\\"\""); |
| 1062 | } |
| 1063 | |
| 1064 | #[test] |
| 1065 | fn roundtrip() { |
| 1066 | fn roundtrip(p: &str) { |
| 1067 | println!("parse: {}", p); |
| 1068 | let s = p.parse::<TokenStream>().unwrap().to_string(); |
| 1069 | println!("first: {}", s); |
| 1070 | let s2 = s.to_string().parse::<TokenStream>().unwrap().to_string(); |
| 1071 | assert_eq!(s, s2); |
| 1072 | } |
| 1073 | roundtrip("a"); |
| 1074 | roundtrip("<<"); |
| 1075 | roundtrip("<<="); |
| 1076 | roundtrip(" |
| 1077 | /// a |
| 1078 | wut |
| 1079 | "); |
| 1080 | roundtrip(" |
| 1081 | 1 |
| 1082 | 1.0 |
| 1083 | 1f32 |
| 1084 | 2f64 |
| 1085 | 1usize |
| 1086 | 4isize |
| 1087 | 4e10 |
| 1088 | 1_000 |
| 1089 | 1_0i32 |
| 1090 | 8u8 |
| 1091 | 9 |
| 1092 | 0 |
| 1093 | "); |
| 1094 | } |
| 1095 | } |