1 | #[cfg (feature = "parsing" )] |
2 | use crate::lookahead; |
3 | #[cfg (feature = "parsing" )] |
4 | use crate::parse::{Parse, Parser}; |
5 | use crate::{Error, Result}; |
6 | use proc_macro2::{Ident, Literal, Span}; |
7 | #[cfg (feature = "parsing" )] |
8 | use proc_macro2::{TokenStream, TokenTree}; |
9 | use std::fmt::{self, Display}; |
10 | #[cfg (feature = "extra-traits" )] |
11 | use std::hash::{Hash, Hasher}; |
12 | use std::str::{self, FromStr}; |
13 | |
14 | ast_enum_of_structs! { |
15 | /// A Rust literal such as a string or integer or boolean. |
16 | /// |
17 | /// # Syntax tree enum |
18 | /// |
19 | /// This type is a [syntax tree enum]. |
20 | /// |
21 | /// [syntax tree enum]: crate::expr::Expr#syntax-tree-enums |
22 | #[non_exhaustive ] |
23 | pub enum Lit { |
24 | /// A UTF-8 string literal: `"foo"`. |
25 | Str(LitStr), |
26 | |
27 | /// A byte string literal: `b"foo"`. |
28 | ByteStr(LitByteStr), |
29 | |
30 | /// A byte literal: `b'f'`. |
31 | Byte(LitByte), |
32 | |
33 | /// A character literal: `'a'`. |
34 | Char(LitChar), |
35 | |
36 | /// An integer literal: `1` or `1u16`. |
37 | Int(LitInt), |
38 | |
39 | /// A floating point literal: `1f64` or `1.0e10f64`. |
40 | /// |
41 | /// Must be finite. May not be infinite or NaN. |
42 | Float(LitFloat), |
43 | |
44 | /// A boolean literal: `true` or `false`. |
45 | Bool(LitBool), |
46 | |
47 | /// A raw token literal not interpreted by Syn. |
48 | Verbatim(Literal), |
49 | } |
50 | } |
51 | |
52 | ast_struct! { |
53 | /// A UTF-8 string literal: `"foo"`. |
54 | pub struct LitStr { |
55 | repr: Box<LitRepr>, |
56 | } |
57 | } |
58 | |
59 | ast_struct! { |
60 | /// A byte string literal: `b"foo"`. |
61 | pub struct LitByteStr { |
62 | repr: Box<LitRepr>, |
63 | } |
64 | } |
65 | |
66 | ast_struct! { |
67 | /// A byte literal: `b'f'`. |
68 | pub struct LitByte { |
69 | repr: Box<LitRepr>, |
70 | } |
71 | } |
72 | |
73 | ast_struct! { |
74 | /// A character literal: `'a'`. |
75 | pub struct LitChar { |
76 | repr: Box<LitRepr>, |
77 | } |
78 | } |
79 | |
80 | struct LitRepr { |
81 | token: Literal, |
82 | suffix: Box<str>, |
83 | } |
84 | |
85 | ast_struct! { |
86 | /// An integer literal: `1` or `1u16`. |
87 | pub struct LitInt { |
88 | repr: Box<LitIntRepr>, |
89 | } |
90 | } |
91 | |
92 | struct LitIntRepr { |
93 | token: Literal, |
94 | digits: Box<str>, |
95 | suffix: Box<str>, |
96 | } |
97 | |
98 | ast_struct! { |
99 | /// A floating point literal: `1f64` or `1.0e10f64`. |
100 | /// |
101 | /// Must be finite. May not be infinite or NaN. |
102 | pub struct LitFloat { |
103 | repr: Box<LitFloatRepr>, |
104 | } |
105 | } |
106 | |
107 | struct LitFloatRepr { |
108 | token: Literal, |
109 | digits: Box<str>, |
110 | suffix: Box<str>, |
111 | } |
112 | |
113 | ast_struct! { |
114 | /// A boolean literal: `true` or `false`. |
115 | pub struct LitBool { |
116 | pub value: bool, |
117 | pub span: Span, |
118 | } |
119 | } |
120 | |
121 | impl LitStr { |
122 | pub fn new(value: &str, span: Span) -> Self { |
123 | let mut token = Literal::string(value); |
124 | token.set_span(span); |
125 | LitStr { |
126 | repr: Box::new(LitRepr { |
127 | token, |
128 | suffix: Box::<str>::default(), |
129 | }), |
130 | } |
131 | } |
132 | |
133 | pub fn value(&self) -> String { |
134 | let repr = self.repr.token.to_string(); |
135 | let (value, _suffix) = value::parse_lit_str(&repr); |
136 | String::from(value) |
137 | } |
138 | |
139 | /// Parse a syntax tree node from the content of this string literal. |
140 | /// |
141 | /// All spans in the syntax tree will point to the span of this `LitStr`. |
142 | /// |
143 | /// # Example |
144 | /// |
145 | /// ``` |
146 | /// use syn::{Attribute, Error, Expr, Lit, Meta, Path, Result}; |
147 | /// |
148 | /// // Parses the path from an attribute that looks like: |
149 | /// // |
150 | /// // #[path = "a::b::c"] |
151 | /// // |
152 | /// // or returns `None` if the input is some other attribute. |
153 | /// fn get_path(attr: &Attribute) -> Result<Option<Path>> { |
154 | /// if !attr.path().is_ident("path" ) { |
155 | /// return Ok(None); |
156 | /// } |
157 | /// |
158 | /// if let Meta::NameValue(meta) = &attr.meta { |
159 | /// if let Expr::Lit(expr) = &meta.value { |
160 | /// if let Lit::Str(lit_str) = &expr.lit { |
161 | /// return lit_str.parse().map(Some); |
162 | /// } |
163 | /// } |
164 | /// } |
165 | /// |
166 | /// let message = "expected #[path = \"... \"]" ; |
167 | /// Err(Error::new_spanned(attr, message)) |
168 | /// } |
169 | /// ``` |
170 | #[cfg (feature = "parsing" )] |
171 | #[cfg_attr (doc_cfg, doc(cfg(feature = "parsing" )))] |
172 | pub fn parse<T: Parse>(&self) -> Result<T> { |
173 | self.parse_with(T::parse) |
174 | } |
175 | |
176 | /// Invoke parser on the content of this string literal. |
177 | /// |
178 | /// All spans in the syntax tree will point to the span of this `LitStr`. |
179 | /// |
180 | /// # Example |
181 | /// |
182 | /// ``` |
183 | /// # use proc_macro2::Span; |
184 | /// # use syn::{LitStr, Result}; |
185 | /// # |
186 | /// # fn main() -> Result<()> { |
187 | /// # let lit_str = LitStr::new("a::b::c" , Span::call_site()); |
188 | /// # |
189 | /// # const IGNORE: &str = stringify! { |
190 | /// let lit_str: LitStr = /* ... */; |
191 | /// # }; |
192 | /// |
193 | /// // Parse a string literal like "a::b::c" into a Path, not allowing |
194 | /// // generic arguments on any of the path segments. |
195 | /// let basic_path = lit_str.parse_with(syn::Path::parse_mod_style)?; |
196 | /// # |
197 | /// # Ok(()) |
198 | /// # } |
199 | /// ``` |
200 | #[cfg (feature = "parsing" )] |
201 | #[cfg_attr (doc_cfg, doc(cfg(feature = "parsing" )))] |
202 | pub fn parse_with<F: Parser>(&self, parser: F) -> Result<F::Output> { |
203 | use proc_macro2::Group; |
204 | |
205 | // Token stream with every span replaced by the given one. |
206 | fn respan_token_stream(stream: TokenStream, span: Span) -> TokenStream { |
207 | stream |
208 | .into_iter() |
209 | .map(|token| respan_token_tree(token, span)) |
210 | .collect() |
211 | } |
212 | |
213 | // Token tree with every span replaced by the given one. |
214 | fn respan_token_tree(mut token: TokenTree, span: Span) -> TokenTree { |
215 | match &mut token { |
216 | TokenTree::Group(g) => { |
217 | let stream = respan_token_stream(g.stream(), span); |
218 | *g = Group::new(g.delimiter(), stream); |
219 | g.set_span(span); |
220 | } |
221 | other => other.set_span(span), |
222 | } |
223 | token |
224 | } |
225 | |
226 | // Parse string literal into a token stream with every span equal to the |
227 | // original literal's span. |
228 | let span = self.span(); |
229 | let mut tokens = TokenStream::from_str(&self.value())?; |
230 | tokens = respan_token_stream(tokens, span); |
231 | |
232 | let result = crate::parse::parse_scoped(parser, span, tokens)?; |
233 | |
234 | let suffix = self.suffix(); |
235 | if !suffix.is_empty() { |
236 | return Err(Error::new( |
237 | self.span(), |
238 | format!("unexpected suffix ` {}` on string literal" , suffix), |
239 | )); |
240 | } |
241 | |
242 | Ok(result) |
243 | } |
244 | |
245 | pub fn span(&self) -> Span { |
246 | self.repr.token.span() |
247 | } |
248 | |
249 | pub fn set_span(&mut self, span: Span) { |
250 | self.repr.token.set_span(span); |
251 | } |
252 | |
253 | pub fn suffix(&self) -> &str { |
254 | &self.repr.suffix |
255 | } |
256 | |
257 | pub fn token(&self) -> Literal { |
258 | self.repr.token.clone() |
259 | } |
260 | } |
261 | |
262 | impl LitByteStr { |
263 | pub fn new(value: &[u8], span: Span) -> Self { |
264 | let mut token = Literal::byte_string(value); |
265 | token.set_span(span); |
266 | LitByteStr { |
267 | repr: Box::new(LitRepr { |
268 | token, |
269 | suffix: Box::<str>::default(), |
270 | }), |
271 | } |
272 | } |
273 | |
274 | pub fn value(&self) -> Vec<u8> { |
275 | let repr = self.repr.token.to_string(); |
276 | let (value, _suffix) = value::parse_lit_byte_str(&repr); |
277 | value |
278 | } |
279 | |
280 | pub fn span(&self) -> Span { |
281 | self.repr.token.span() |
282 | } |
283 | |
284 | pub fn set_span(&mut self, span: Span) { |
285 | self.repr.token.set_span(span); |
286 | } |
287 | |
288 | pub fn suffix(&self) -> &str { |
289 | &self.repr.suffix |
290 | } |
291 | |
292 | pub fn token(&self) -> Literal { |
293 | self.repr.token.clone() |
294 | } |
295 | } |
296 | |
297 | impl LitByte { |
298 | pub fn new(value: u8, span: Span) -> Self { |
299 | let mut token = Literal::u8_suffixed(value); |
300 | token.set_span(span); |
301 | LitByte { |
302 | repr: Box::new(LitRepr { |
303 | token, |
304 | suffix: Box::<str>::default(), |
305 | }), |
306 | } |
307 | } |
308 | |
309 | pub fn value(&self) -> u8 { |
310 | let repr = self.repr.token.to_string(); |
311 | let (value, _suffix) = value::parse_lit_byte(&repr); |
312 | value |
313 | } |
314 | |
315 | pub fn span(&self) -> Span { |
316 | self.repr.token.span() |
317 | } |
318 | |
319 | pub fn set_span(&mut self, span: Span) { |
320 | self.repr.token.set_span(span); |
321 | } |
322 | |
323 | pub fn suffix(&self) -> &str { |
324 | &self.repr.suffix |
325 | } |
326 | |
327 | pub fn token(&self) -> Literal { |
328 | self.repr.token.clone() |
329 | } |
330 | } |
331 | |
332 | impl LitChar { |
333 | pub fn new(value: char, span: Span) -> Self { |
334 | let mut token = Literal::character(value); |
335 | token.set_span(span); |
336 | LitChar { |
337 | repr: Box::new(LitRepr { |
338 | token, |
339 | suffix: Box::<str>::default(), |
340 | }), |
341 | } |
342 | } |
343 | |
344 | pub fn value(&self) -> char { |
345 | let repr = self.repr.token.to_string(); |
346 | let (value, _suffix) = value::parse_lit_char(&repr); |
347 | value |
348 | } |
349 | |
350 | pub fn span(&self) -> Span { |
351 | self.repr.token.span() |
352 | } |
353 | |
354 | pub fn set_span(&mut self, span: Span) { |
355 | self.repr.token.set_span(span); |
356 | } |
357 | |
358 | pub fn suffix(&self) -> &str { |
359 | &self.repr.suffix |
360 | } |
361 | |
362 | pub fn token(&self) -> Literal { |
363 | self.repr.token.clone() |
364 | } |
365 | } |
366 | |
367 | impl LitInt { |
368 | pub fn new(repr: &str, span: Span) -> Self { |
369 | let (digits, suffix) = match value::parse_lit_int(repr) { |
370 | Some(parse) => parse, |
371 | None => panic!("Not an integer literal: ` {}`" , repr), |
372 | }; |
373 | |
374 | let mut token: Literal = repr.parse().unwrap(); |
375 | token.set_span(span); |
376 | LitInt { |
377 | repr: Box::new(LitIntRepr { |
378 | token, |
379 | digits, |
380 | suffix, |
381 | }), |
382 | } |
383 | } |
384 | |
385 | pub fn base10_digits(&self) -> &str { |
386 | &self.repr.digits |
387 | } |
388 | |
389 | /// Parses the literal into a selected number type. |
390 | /// |
391 | /// This is equivalent to `lit.base10_digits().parse()` except that the |
392 | /// resulting errors will be correctly spanned to point to the literal token |
393 | /// in the macro input. |
394 | /// |
395 | /// ``` |
396 | /// use syn::LitInt; |
397 | /// use syn::parse::{Parse, ParseStream, Result}; |
398 | /// |
399 | /// struct Port { |
400 | /// value: u16, |
401 | /// } |
402 | /// |
403 | /// impl Parse for Port { |
404 | /// fn parse(input: ParseStream) -> Result<Self> { |
405 | /// let lit: LitInt = input.parse()?; |
406 | /// let value = lit.base10_parse::<u16>()?; |
407 | /// Ok(Port { value }) |
408 | /// } |
409 | /// } |
410 | /// ``` |
411 | pub fn base10_parse<N>(&self) -> Result<N> |
412 | where |
413 | N: FromStr, |
414 | N::Err: Display, |
415 | { |
416 | self.base10_digits() |
417 | .parse() |
418 | .map_err(|err| Error::new(self.span(), err)) |
419 | } |
420 | |
421 | pub fn suffix(&self) -> &str { |
422 | &self.repr.suffix |
423 | } |
424 | |
425 | pub fn span(&self) -> Span { |
426 | self.repr.token.span() |
427 | } |
428 | |
429 | pub fn set_span(&mut self, span: Span) { |
430 | self.repr.token.set_span(span); |
431 | } |
432 | |
433 | pub fn token(&self) -> Literal { |
434 | self.repr.token.clone() |
435 | } |
436 | } |
437 | |
438 | impl From<Literal> for LitInt { |
439 | fn from(token: Literal) -> Self { |
440 | let repr: String = token.to_string(); |
441 | if let Some((digits: Box, suffix: Box)) = value::parse_lit_int(&repr) { |
442 | LitInt { |
443 | repr: Box::new(LitIntRepr { |
444 | token, |
445 | digits, |
446 | suffix, |
447 | }), |
448 | } |
449 | } else { |
450 | panic!("Not an integer literal: ` {}`" , repr); |
451 | } |
452 | } |
453 | } |
454 | |
455 | impl Display for LitInt { |
456 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
457 | self.repr.token.fmt(formatter) |
458 | } |
459 | } |
460 | |
461 | impl LitFloat { |
462 | pub fn new(repr: &str, span: Span) -> Self { |
463 | let (digits, suffix) = match value::parse_lit_float(repr) { |
464 | Some(parse) => parse, |
465 | None => panic!("Not a float literal: ` {}`" , repr), |
466 | }; |
467 | |
468 | let mut token: Literal = repr.parse().unwrap(); |
469 | token.set_span(span); |
470 | LitFloat { |
471 | repr: Box::new(LitFloatRepr { |
472 | token, |
473 | digits, |
474 | suffix, |
475 | }), |
476 | } |
477 | } |
478 | |
479 | pub fn base10_digits(&self) -> &str { |
480 | &self.repr.digits |
481 | } |
482 | |
483 | pub fn base10_parse<N>(&self) -> Result<N> |
484 | where |
485 | N: FromStr, |
486 | N::Err: Display, |
487 | { |
488 | self.base10_digits() |
489 | .parse() |
490 | .map_err(|err| Error::new(self.span(), err)) |
491 | } |
492 | |
493 | pub fn suffix(&self) -> &str { |
494 | &self.repr.suffix |
495 | } |
496 | |
497 | pub fn span(&self) -> Span { |
498 | self.repr.token.span() |
499 | } |
500 | |
501 | pub fn set_span(&mut self, span: Span) { |
502 | self.repr.token.set_span(span); |
503 | } |
504 | |
505 | pub fn token(&self) -> Literal { |
506 | self.repr.token.clone() |
507 | } |
508 | } |
509 | |
510 | impl From<Literal> for LitFloat { |
511 | fn from(token: Literal) -> Self { |
512 | let repr: String = token.to_string(); |
513 | if let Some((digits: Box, suffix: Box)) = value::parse_lit_float(&repr) { |
514 | LitFloat { |
515 | repr: Box::new(LitFloatRepr { |
516 | token, |
517 | digits, |
518 | suffix, |
519 | }), |
520 | } |
521 | } else { |
522 | panic!("Not a float literal: ` {}`" , repr); |
523 | } |
524 | } |
525 | } |
526 | |
527 | impl Display for LitFloat { |
528 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
529 | self.repr.token.fmt(formatter) |
530 | } |
531 | } |
532 | |
533 | impl LitBool { |
534 | pub fn new(value: bool, span: Span) -> Self { |
535 | LitBool { value, span } |
536 | } |
537 | |
538 | pub fn value(&self) -> bool { |
539 | self.value |
540 | } |
541 | |
542 | pub fn span(&self) -> Span { |
543 | self.span |
544 | } |
545 | |
546 | pub fn set_span(&mut self, span: Span) { |
547 | self.span = span; |
548 | } |
549 | |
550 | pub fn token(&self) -> Ident { |
551 | let s: &str = if self.value { "true" } else { "false" }; |
552 | Ident::new(string:s, self.span) |
553 | } |
554 | } |
555 | |
556 | #[cfg (feature = "extra-traits" )] |
557 | mod debug_impls { |
558 | use crate::lit::{LitBool, LitByte, LitByteStr, LitChar, LitFloat, LitInt, LitStr}; |
559 | use std::fmt::{self, Debug}; |
560 | |
561 | #[cfg_attr (doc_cfg, doc(cfg(feature = "extra-traits" )))] |
562 | impl Debug for LitStr { |
563 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
564 | self.debug(formatter, "LitStr" ) |
565 | } |
566 | } |
567 | |
568 | impl LitStr { |
569 | pub(crate) fn debug(&self, formatter: &mut fmt::Formatter, name: &str) -> fmt::Result { |
570 | formatter |
571 | .debug_struct(name) |
572 | .field("token" , &format_args!(" {}" , self.repr.token)) |
573 | .finish() |
574 | } |
575 | } |
576 | |
577 | #[cfg_attr (doc_cfg, doc(cfg(feature = "extra-traits" )))] |
578 | impl Debug for LitByteStr { |
579 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
580 | self.debug(formatter, "LitByteStr" ) |
581 | } |
582 | } |
583 | |
584 | impl LitByteStr { |
585 | pub(crate) fn debug(&self, formatter: &mut fmt::Formatter, name: &str) -> fmt::Result { |
586 | formatter |
587 | .debug_struct(name) |
588 | .field("token" , &format_args!(" {}" , self.repr.token)) |
589 | .finish() |
590 | } |
591 | } |
592 | |
593 | #[cfg_attr (doc_cfg, doc(cfg(feature = "extra-traits" )))] |
594 | impl Debug for LitByte { |
595 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
596 | self.debug(formatter, "LitByte" ) |
597 | } |
598 | } |
599 | |
600 | impl LitByte { |
601 | pub(crate) fn debug(&self, formatter: &mut fmt::Formatter, name: &str) -> fmt::Result { |
602 | formatter |
603 | .debug_struct(name) |
604 | .field("token" , &format_args!(" {}" , self.repr.token)) |
605 | .finish() |
606 | } |
607 | } |
608 | |
609 | #[cfg_attr (doc_cfg, doc(cfg(feature = "extra-traits" )))] |
610 | impl Debug for LitChar { |
611 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
612 | self.debug(formatter, "LitChar" ) |
613 | } |
614 | } |
615 | |
616 | impl LitChar { |
617 | pub(crate) fn debug(&self, formatter: &mut fmt::Formatter, name: &str) -> fmt::Result { |
618 | formatter |
619 | .debug_struct(name) |
620 | .field("token" , &format_args!(" {}" , self.repr.token)) |
621 | .finish() |
622 | } |
623 | } |
624 | |
625 | #[cfg_attr (doc_cfg, doc(cfg(feature = "extra-traits" )))] |
626 | impl Debug for LitInt { |
627 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
628 | self.debug(formatter, "LitInt" ) |
629 | } |
630 | } |
631 | |
632 | impl LitInt { |
633 | pub(crate) fn debug(&self, formatter: &mut fmt::Formatter, name: &str) -> fmt::Result { |
634 | formatter |
635 | .debug_struct(name) |
636 | .field("token" , &format_args!(" {}" , self.repr.token)) |
637 | .finish() |
638 | } |
639 | } |
640 | |
641 | #[cfg_attr (doc_cfg, doc(cfg(feature = "extra-traits" )))] |
642 | impl Debug for LitFloat { |
643 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
644 | self.debug(formatter, "LitFloat" ) |
645 | } |
646 | } |
647 | |
648 | impl LitFloat { |
649 | pub(crate) fn debug(&self, formatter: &mut fmt::Formatter, name: &str) -> fmt::Result { |
650 | formatter |
651 | .debug_struct(name) |
652 | .field("token" , &format_args!(" {}" , self.repr.token)) |
653 | .finish() |
654 | } |
655 | } |
656 | |
657 | #[cfg_attr (doc_cfg, doc(cfg(feature = "extra-traits" )))] |
658 | impl Debug for LitBool { |
659 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
660 | self.debug(formatter, "LitBool" ) |
661 | } |
662 | } |
663 | |
664 | impl LitBool { |
665 | pub(crate) fn debug(&self, formatter: &mut fmt::Formatter, name: &str) -> fmt::Result { |
666 | formatter |
667 | .debug_struct(name) |
668 | .field("value" , &self.value) |
669 | .finish() |
670 | } |
671 | } |
672 | } |
673 | |
674 | #[cfg (feature = "clone-impls" )] |
675 | #[cfg_attr (doc_cfg, doc(cfg(feature = "clone-impls" )))] |
676 | impl Clone for LitRepr { |
677 | fn clone(&self) -> Self { |
678 | LitRepr { |
679 | token: self.token.clone(), |
680 | suffix: self.suffix.clone(), |
681 | } |
682 | } |
683 | } |
684 | |
685 | #[cfg (feature = "clone-impls" )] |
686 | #[cfg_attr (doc_cfg, doc(cfg(feature = "clone-impls" )))] |
687 | impl Clone for LitIntRepr { |
688 | fn clone(&self) -> Self { |
689 | LitIntRepr { |
690 | token: self.token.clone(), |
691 | digits: self.digits.clone(), |
692 | suffix: self.suffix.clone(), |
693 | } |
694 | } |
695 | } |
696 | |
697 | #[cfg (feature = "clone-impls" )] |
698 | #[cfg_attr (doc_cfg, doc(cfg(feature = "clone-impls" )))] |
699 | impl Clone for LitFloatRepr { |
700 | fn clone(&self) -> Self { |
701 | LitFloatRepr { |
702 | token: self.token.clone(), |
703 | digits: self.digits.clone(), |
704 | suffix: self.suffix.clone(), |
705 | } |
706 | } |
707 | } |
708 | |
709 | macro_rules! lit_extra_traits { |
710 | ($ty:ident) => { |
711 | #[cfg(feature = "clone-impls" )] |
712 | #[cfg_attr(doc_cfg, doc(cfg(feature = "clone-impls" )))] |
713 | impl Clone for $ty { |
714 | fn clone(&self) -> Self { |
715 | $ty { |
716 | repr: self.repr.clone(), |
717 | } |
718 | } |
719 | } |
720 | |
721 | #[cfg(feature = "extra-traits" )] |
722 | #[cfg_attr(doc_cfg, doc(cfg(feature = "extra-traits" )))] |
723 | impl PartialEq for $ty { |
724 | fn eq(&self, other: &Self) -> bool { |
725 | self.repr.token.to_string() == other.repr.token.to_string() |
726 | } |
727 | } |
728 | |
729 | #[cfg(feature = "extra-traits" )] |
730 | #[cfg_attr(doc_cfg, doc(cfg(feature = "extra-traits" )))] |
731 | impl Hash for $ty { |
732 | fn hash<H>(&self, state: &mut H) |
733 | where |
734 | H: Hasher, |
735 | { |
736 | self.repr.token.to_string().hash(state); |
737 | } |
738 | } |
739 | |
740 | #[cfg(feature = "parsing" )] |
741 | pub_if_not_doc! { |
742 | #[doc(hidden)] |
743 | #[allow(non_snake_case)] |
744 | pub fn $ty(marker: lookahead::TokenMarker) -> $ty { |
745 | match marker {} |
746 | } |
747 | } |
748 | }; |
749 | } |
750 | |
751 | lit_extra_traits!(LitStr); |
752 | lit_extra_traits!(LitByteStr); |
753 | lit_extra_traits!(LitByte); |
754 | lit_extra_traits!(LitChar); |
755 | lit_extra_traits!(LitInt); |
756 | lit_extra_traits!(LitFloat); |
757 | |
758 | #[cfg (feature = "parsing" )] |
759 | pub_if_not_doc! { |
760 | #[doc (hidden)] |
761 | #[allow (non_snake_case)] |
762 | pub fn LitBool(marker: lookahead::TokenMarker) -> LitBool { |
763 | match marker {} |
764 | } |
765 | } |
766 | |
767 | /// The style of a string literal, either plain quoted or a raw string like |
768 | /// `r##"data"##`. |
769 | #[doc (hidden)] // https://github.com/dtolnay/syn/issues/1566 |
770 | pub enum StrStyle { |
771 | /// An ordinary string like `"data"`. |
772 | Cooked, |
773 | /// A raw string like `r##"data"##`. |
774 | /// |
775 | /// The unsigned integer is the number of `#` symbols used. |
776 | Raw(usize), |
777 | } |
778 | |
779 | #[cfg (feature = "parsing" )] |
780 | pub_if_not_doc! { |
781 | #[doc (hidden)] |
782 | #[allow (non_snake_case)] |
783 | pub fn Lit(marker: lookahead::TokenMarker) -> Lit { |
784 | match marker {} |
785 | } |
786 | } |
787 | |
788 | #[cfg (feature = "parsing" )] |
789 | pub(crate) mod parsing { |
790 | use crate::buffer::Cursor; |
791 | use crate::error::Result; |
792 | use crate::lit::{ |
793 | value, Lit, LitBool, LitByte, LitByteStr, LitChar, LitFloat, LitFloatRepr, LitInt, |
794 | LitIntRepr, LitStr, |
795 | }; |
796 | use crate::parse::{Parse, ParseStream}; |
797 | use proc_macro2::{Literal, Punct}; |
798 | |
799 | #[cfg_attr (doc_cfg, doc(cfg(feature = "parsing" )))] |
800 | impl Parse for Lit { |
801 | fn parse(input: ParseStream) -> Result<Self> { |
802 | input.step(|cursor| { |
803 | if let Some((lit, rest)) = cursor.literal() { |
804 | return Ok((Lit::new(lit), rest)); |
805 | } |
806 | |
807 | if let Some((ident, rest)) = cursor.ident() { |
808 | let value = ident == "true" ; |
809 | if value || ident == "false" { |
810 | let lit_bool = LitBool { |
811 | value, |
812 | span: ident.span(), |
813 | }; |
814 | return Ok((Lit::Bool(lit_bool), rest)); |
815 | } |
816 | } |
817 | |
818 | if let Some((punct, rest)) = cursor.punct() { |
819 | if punct.as_char() == '-' { |
820 | if let Some((lit, rest)) = parse_negative_lit(punct, rest) { |
821 | return Ok((lit, rest)); |
822 | } |
823 | } |
824 | } |
825 | |
826 | Err(cursor.error("expected literal" )) |
827 | }) |
828 | } |
829 | } |
830 | |
831 | fn parse_negative_lit(neg: Punct, cursor: Cursor) -> Option<(Lit, Cursor)> { |
832 | let (lit, rest) = cursor.literal()?; |
833 | |
834 | let mut span = neg.span(); |
835 | span = span.join(lit.span()).unwrap_or(span); |
836 | |
837 | let mut repr = lit.to_string(); |
838 | repr.insert(0, '-' ); |
839 | |
840 | if let Some((digits, suffix)) = value::parse_lit_int(&repr) { |
841 | let mut token: Literal = repr.parse().unwrap(); |
842 | token.set_span(span); |
843 | return Some(( |
844 | Lit::Int(LitInt { |
845 | repr: Box::new(LitIntRepr { |
846 | token, |
847 | digits, |
848 | suffix, |
849 | }), |
850 | }), |
851 | rest, |
852 | )); |
853 | } |
854 | |
855 | let (digits, suffix) = value::parse_lit_float(&repr)?; |
856 | let mut token: Literal = repr.parse().unwrap(); |
857 | token.set_span(span); |
858 | Some(( |
859 | Lit::Float(LitFloat { |
860 | repr: Box::new(LitFloatRepr { |
861 | token, |
862 | digits, |
863 | suffix, |
864 | }), |
865 | }), |
866 | rest, |
867 | )) |
868 | } |
869 | |
870 | #[cfg_attr (doc_cfg, doc(cfg(feature = "parsing" )))] |
871 | impl Parse for LitStr { |
872 | fn parse(input: ParseStream) -> Result<Self> { |
873 | let head = input.fork(); |
874 | match input.parse() { |
875 | Ok(Lit::Str(lit)) => Ok(lit), |
876 | _ => Err(head.error("expected string literal" )), |
877 | } |
878 | } |
879 | } |
880 | |
881 | #[cfg_attr (doc_cfg, doc(cfg(feature = "parsing" )))] |
882 | impl Parse for LitByteStr { |
883 | fn parse(input: ParseStream) -> Result<Self> { |
884 | let head = input.fork(); |
885 | match input.parse() { |
886 | Ok(Lit::ByteStr(lit)) => Ok(lit), |
887 | _ => Err(head.error("expected byte string literal" )), |
888 | } |
889 | } |
890 | } |
891 | |
892 | #[cfg_attr (doc_cfg, doc(cfg(feature = "parsing" )))] |
893 | impl Parse for LitByte { |
894 | fn parse(input: ParseStream) -> Result<Self> { |
895 | let head = input.fork(); |
896 | match input.parse() { |
897 | Ok(Lit::Byte(lit)) => Ok(lit), |
898 | _ => Err(head.error("expected byte literal" )), |
899 | } |
900 | } |
901 | } |
902 | |
903 | #[cfg_attr (doc_cfg, doc(cfg(feature = "parsing" )))] |
904 | impl Parse for LitChar { |
905 | fn parse(input: ParseStream) -> Result<Self> { |
906 | let head = input.fork(); |
907 | match input.parse() { |
908 | Ok(Lit::Char(lit)) => Ok(lit), |
909 | _ => Err(head.error("expected character literal" )), |
910 | } |
911 | } |
912 | } |
913 | |
914 | #[cfg_attr (doc_cfg, doc(cfg(feature = "parsing" )))] |
915 | impl Parse for LitInt { |
916 | fn parse(input: ParseStream) -> Result<Self> { |
917 | let head = input.fork(); |
918 | match input.parse() { |
919 | Ok(Lit::Int(lit)) => Ok(lit), |
920 | _ => Err(head.error("expected integer literal" )), |
921 | } |
922 | } |
923 | } |
924 | |
925 | #[cfg_attr (doc_cfg, doc(cfg(feature = "parsing" )))] |
926 | impl Parse for LitFloat { |
927 | fn parse(input: ParseStream) -> Result<Self> { |
928 | let head = input.fork(); |
929 | match input.parse() { |
930 | Ok(Lit::Float(lit)) => Ok(lit), |
931 | _ => Err(head.error("expected floating point literal" )), |
932 | } |
933 | } |
934 | } |
935 | |
936 | #[cfg_attr (doc_cfg, doc(cfg(feature = "parsing" )))] |
937 | impl Parse for LitBool { |
938 | fn parse(input: ParseStream) -> Result<Self> { |
939 | let head = input.fork(); |
940 | match input.parse() { |
941 | Ok(Lit::Bool(lit)) => Ok(lit), |
942 | _ => Err(head.error("expected boolean literal" )), |
943 | } |
944 | } |
945 | } |
946 | } |
947 | |
948 | #[cfg (feature = "printing" )] |
949 | mod printing { |
950 | use crate::lit::{LitBool, LitByte, LitByteStr, LitChar, LitFloat, LitInt, LitStr}; |
951 | use proc_macro2::TokenStream; |
952 | use quote::{ToTokens, TokenStreamExt}; |
953 | |
954 | #[cfg_attr (doc_cfg, doc(cfg(feature = "printing" )))] |
955 | impl ToTokens for LitStr { |
956 | fn to_tokens(&self, tokens: &mut TokenStream) { |
957 | self.repr.token.to_tokens(tokens); |
958 | } |
959 | } |
960 | |
961 | #[cfg_attr (doc_cfg, doc(cfg(feature = "printing" )))] |
962 | impl ToTokens for LitByteStr { |
963 | fn to_tokens(&self, tokens: &mut TokenStream) { |
964 | self.repr.token.to_tokens(tokens); |
965 | } |
966 | } |
967 | |
968 | #[cfg_attr (doc_cfg, doc(cfg(feature = "printing" )))] |
969 | impl ToTokens for LitByte { |
970 | fn to_tokens(&self, tokens: &mut TokenStream) { |
971 | self.repr.token.to_tokens(tokens); |
972 | } |
973 | } |
974 | |
975 | #[cfg_attr (doc_cfg, doc(cfg(feature = "printing" )))] |
976 | impl ToTokens for LitChar { |
977 | fn to_tokens(&self, tokens: &mut TokenStream) { |
978 | self.repr.token.to_tokens(tokens); |
979 | } |
980 | } |
981 | |
982 | #[cfg_attr (doc_cfg, doc(cfg(feature = "printing" )))] |
983 | impl ToTokens for LitInt { |
984 | fn to_tokens(&self, tokens: &mut TokenStream) { |
985 | self.repr.token.to_tokens(tokens); |
986 | } |
987 | } |
988 | |
989 | #[cfg_attr (doc_cfg, doc(cfg(feature = "printing" )))] |
990 | impl ToTokens for LitFloat { |
991 | fn to_tokens(&self, tokens: &mut TokenStream) { |
992 | self.repr.token.to_tokens(tokens); |
993 | } |
994 | } |
995 | |
996 | #[cfg_attr (doc_cfg, doc(cfg(feature = "printing" )))] |
997 | impl ToTokens for LitBool { |
998 | fn to_tokens(&self, tokens: &mut TokenStream) { |
999 | tokens.append(self.token()); |
1000 | } |
1001 | } |
1002 | } |
1003 | |
1004 | mod value { |
1005 | use crate::bigint::BigInt; |
1006 | use crate::lit::{ |
1007 | Lit, LitBool, LitByte, LitByteStr, LitChar, LitFloat, LitFloatRepr, LitInt, LitIntRepr, |
1008 | LitRepr, LitStr, |
1009 | }; |
1010 | use proc_macro2::{Literal, Span}; |
1011 | use std::char; |
1012 | use std::ops::{Index, RangeFrom}; |
1013 | |
1014 | impl Lit { |
1015 | /// Interpret a Syn literal from a proc-macro2 literal. |
1016 | pub fn new(token: Literal) -> Self { |
1017 | let repr = token.to_string(); |
1018 | |
1019 | match byte(&repr, 0) { |
1020 | // "...", r"...", r#"..."# |
1021 | b'"' | b'r' => { |
1022 | let (_, suffix) = parse_lit_str(&repr); |
1023 | return Lit::Str(LitStr { |
1024 | repr: Box::new(LitRepr { token, suffix }), |
1025 | }); |
1026 | } |
1027 | b'b' => match byte(&repr, 1) { |
1028 | // b"...", br"...", br#"...#" |
1029 | b'"' | b'r' => { |
1030 | let (_, suffix) = parse_lit_byte_str(&repr); |
1031 | return Lit::ByteStr(LitByteStr { |
1032 | repr: Box::new(LitRepr { token, suffix }), |
1033 | }); |
1034 | } |
1035 | // b'...' |
1036 | b' \'' => { |
1037 | let (_, suffix) = parse_lit_byte(&repr); |
1038 | return Lit::Byte(LitByte { |
1039 | repr: Box::new(LitRepr { token, suffix }), |
1040 | }); |
1041 | } |
1042 | _ => {} |
1043 | }, |
1044 | // '...' |
1045 | b' \'' => { |
1046 | let (_, suffix) = parse_lit_char(&repr); |
1047 | return Lit::Char(LitChar { |
1048 | repr: Box::new(LitRepr { token, suffix }), |
1049 | }); |
1050 | } |
1051 | b'0' ..=b'9' | b'-' => { |
1052 | // 0, 123, 0xFF, 0o77, 0b11 |
1053 | if let Some((digits, suffix)) = parse_lit_int(&repr) { |
1054 | return Lit::Int(LitInt { |
1055 | repr: Box::new(LitIntRepr { |
1056 | token, |
1057 | digits, |
1058 | suffix, |
1059 | }), |
1060 | }); |
1061 | } |
1062 | // 1.0, 1e-1, 1e+1 |
1063 | if let Some((digits, suffix)) = parse_lit_float(&repr) { |
1064 | return Lit::Float(LitFloat { |
1065 | repr: Box::new(LitFloatRepr { |
1066 | token, |
1067 | digits, |
1068 | suffix, |
1069 | }), |
1070 | }); |
1071 | } |
1072 | } |
1073 | // true, false |
1074 | b't' | b'f' => { |
1075 | if repr == "true" || repr == "false" { |
1076 | return Lit::Bool(LitBool { |
1077 | value: repr == "true" , |
1078 | span: token.span(), |
1079 | }); |
1080 | } |
1081 | } |
1082 | // c"...", cr"...", cr#"..."# |
1083 | // TODO: add a Lit::CStr variant? |
1084 | b'c' => return Lit::Verbatim(token), |
1085 | b'(' if repr == "(/*ERROR*/)" => return Lit::Verbatim(token), |
1086 | _ => {} |
1087 | } |
1088 | |
1089 | panic!("Unrecognized literal: ` {}`" , repr); |
1090 | } |
1091 | |
1092 | pub fn suffix(&self) -> &str { |
1093 | match self { |
1094 | Lit::Str(lit) => lit.suffix(), |
1095 | Lit::ByteStr(lit) => lit.suffix(), |
1096 | Lit::Byte(lit) => lit.suffix(), |
1097 | Lit::Char(lit) => lit.suffix(), |
1098 | Lit::Int(lit) => lit.suffix(), |
1099 | Lit::Float(lit) => lit.suffix(), |
1100 | Lit::Bool(_) | Lit::Verbatim(_) => "" , |
1101 | } |
1102 | } |
1103 | |
1104 | pub fn span(&self) -> Span { |
1105 | match self { |
1106 | Lit::Str(lit) => lit.span(), |
1107 | Lit::ByteStr(lit) => lit.span(), |
1108 | Lit::Byte(lit) => lit.span(), |
1109 | Lit::Char(lit) => lit.span(), |
1110 | Lit::Int(lit) => lit.span(), |
1111 | Lit::Float(lit) => lit.span(), |
1112 | Lit::Bool(lit) => lit.span, |
1113 | Lit::Verbatim(lit) => lit.span(), |
1114 | } |
1115 | } |
1116 | |
1117 | pub fn set_span(&mut self, span: Span) { |
1118 | match self { |
1119 | Lit::Str(lit) => lit.set_span(span), |
1120 | Lit::ByteStr(lit) => lit.set_span(span), |
1121 | Lit::Byte(lit) => lit.set_span(span), |
1122 | Lit::Char(lit) => lit.set_span(span), |
1123 | Lit::Int(lit) => lit.set_span(span), |
1124 | Lit::Float(lit) => lit.set_span(span), |
1125 | Lit::Bool(lit) => lit.span = span, |
1126 | Lit::Verbatim(lit) => lit.set_span(span), |
1127 | } |
1128 | } |
1129 | } |
1130 | |
1131 | /// Get the byte at offset idx, or a default of `b'\0'` if we're looking |
1132 | /// past the end of the input buffer. |
1133 | pub(crate) fn byte<S: AsRef<[u8]> + ?Sized>(s: &S, idx: usize) -> u8 { |
1134 | let s = s.as_ref(); |
1135 | if idx < s.len() { |
1136 | s[idx] |
1137 | } else { |
1138 | 0 |
1139 | } |
1140 | } |
1141 | |
1142 | fn next_chr(s: &str) -> char { |
1143 | s.chars().next().unwrap_or(' \0' ) |
1144 | } |
1145 | |
1146 | // Returns (content, suffix). |
1147 | pub(crate) fn parse_lit_str(s: &str) -> (Box<str>, Box<str>) { |
1148 | match byte(s, 0) { |
1149 | b'"' => parse_lit_str_cooked(s), |
1150 | b'r' => parse_lit_str_raw(s), |
1151 | _ => unreachable!(), |
1152 | } |
1153 | } |
1154 | |
1155 | // Clippy false positive |
1156 | // https://github.com/rust-lang-nursery/rust-clippy/issues/2329 |
1157 | #[allow (clippy::needless_continue)] |
1158 | fn parse_lit_str_cooked(mut s: &str) -> (Box<str>, Box<str>) { |
1159 | assert_eq!(byte(s, 0), b'"' ); |
1160 | s = &s[1..]; |
1161 | |
1162 | let mut content = String::new(); |
1163 | 'outer: loop { |
1164 | let ch = match byte(s, 0) { |
1165 | b'"' => break, |
1166 | b' \\' => { |
1167 | let b = byte(s, 1); |
1168 | s = &s[2..]; |
1169 | match b { |
1170 | b'x' => { |
1171 | let (byte, rest) = backslash_x(s); |
1172 | s = rest; |
1173 | assert!(byte <= 0x7F, "Invalid \\x byte in string literal" ); |
1174 | char::from_u32(u32::from(byte)).unwrap() |
1175 | } |
1176 | b'u' => { |
1177 | let (chr, rest) = backslash_u(s); |
1178 | s = rest; |
1179 | chr |
1180 | } |
1181 | b'n' => ' \n' , |
1182 | b'r' => ' \r' , |
1183 | b't' => ' \t' , |
1184 | b' \\' => ' \\' , |
1185 | b'0' => ' \0' , |
1186 | b' \'' => ' \'' , |
1187 | b'"' => '"' , |
1188 | b' \r' | b' \n' => loop { |
1189 | let b = byte(s, 0); |
1190 | match b { |
1191 | b' ' | b' \t' | b' \n' | b' \r' => s = &s[1..], |
1192 | _ => continue 'outer, |
1193 | } |
1194 | }, |
1195 | b => panic!("unexpected byte {:?} after \\ character in byte literal" , b), |
1196 | } |
1197 | } |
1198 | b' \r' => { |
1199 | assert_eq!(byte(s, 1), b' \n' , "Bare CR not allowed in string" ); |
1200 | s = &s[2..]; |
1201 | ' \n' |
1202 | } |
1203 | _ => { |
1204 | let ch = next_chr(s); |
1205 | s = &s[ch.len_utf8()..]; |
1206 | ch |
1207 | } |
1208 | }; |
1209 | content.push(ch); |
1210 | } |
1211 | |
1212 | assert!(s.starts_with('"' )); |
1213 | let content = content.into_boxed_str(); |
1214 | let suffix = s[1..].to_owned().into_boxed_str(); |
1215 | (content, suffix) |
1216 | } |
1217 | |
1218 | fn parse_lit_str_raw(mut s: &str) -> (Box<str>, Box<str>) { |
1219 | assert_eq!(byte(s, 0), b'r' ); |
1220 | s = &s[1..]; |
1221 | |
1222 | let mut pounds = 0; |
1223 | while byte(s, pounds) == b'#' { |
1224 | pounds += 1; |
1225 | } |
1226 | assert_eq!(byte(s, pounds), b'"' ); |
1227 | let close = s.rfind('"' ).unwrap(); |
1228 | for end in s[close + 1..close + 1 + pounds].bytes() { |
1229 | assert_eq!(end, b'#' ); |
1230 | } |
1231 | |
1232 | let content = s[pounds + 1..close].to_owned().into_boxed_str(); |
1233 | let suffix = s[close + 1 + pounds..].to_owned().into_boxed_str(); |
1234 | (content, suffix) |
1235 | } |
1236 | |
1237 | // Returns (content, suffix). |
1238 | pub(crate) fn parse_lit_byte_str(s: &str) -> (Vec<u8>, Box<str>) { |
1239 | assert_eq!(byte(s, 0), b'b' ); |
1240 | match byte(s, 1) { |
1241 | b'"' => parse_lit_byte_str_cooked(s), |
1242 | b'r' => parse_lit_byte_str_raw(s), |
1243 | _ => unreachable!(), |
1244 | } |
1245 | } |
1246 | |
1247 | // Clippy false positive |
1248 | // https://github.com/rust-lang-nursery/rust-clippy/issues/2329 |
1249 | #[allow (clippy::needless_continue)] |
1250 | fn parse_lit_byte_str_cooked(mut s: &str) -> (Vec<u8>, Box<str>) { |
1251 | assert_eq!(byte(s, 0), b'b' ); |
1252 | assert_eq!(byte(s, 1), b'"' ); |
1253 | s = &s[2..]; |
1254 | |
1255 | // We're going to want to have slices which don't respect codepoint boundaries. |
1256 | let mut v = s.as_bytes(); |
1257 | |
1258 | let mut out = Vec::new(); |
1259 | 'outer: loop { |
1260 | let byte = match byte(v, 0) { |
1261 | b'"' => break, |
1262 | b' \\' => { |
1263 | let b = byte(v, 1); |
1264 | v = &v[2..]; |
1265 | match b { |
1266 | b'x' => { |
1267 | let (b, rest) = backslash_x(v); |
1268 | v = rest; |
1269 | b |
1270 | } |
1271 | b'n' => b' \n' , |
1272 | b'r' => b' \r' , |
1273 | b't' => b' \t' , |
1274 | b' \\' => b' \\' , |
1275 | b'0' => b' \0' , |
1276 | b' \'' => b' \'' , |
1277 | b'"' => b'"' , |
1278 | b' \r' | b' \n' => loop { |
1279 | let byte = byte(v, 0); |
1280 | if matches!(byte, b' ' | b' \t' | b' \n' | b' \r' ) { |
1281 | v = &v[1..]; |
1282 | } else { |
1283 | continue 'outer; |
1284 | } |
1285 | }, |
1286 | b => panic!("unexpected byte {:?} after \\ character in byte literal" , b), |
1287 | } |
1288 | } |
1289 | b' \r' => { |
1290 | assert_eq!(byte(v, 1), b' \n' , "Bare CR not allowed in string" ); |
1291 | v = &v[2..]; |
1292 | b' \n' |
1293 | } |
1294 | b => { |
1295 | v = &v[1..]; |
1296 | b |
1297 | } |
1298 | }; |
1299 | out.push(byte); |
1300 | } |
1301 | |
1302 | assert_eq!(byte(v, 0), b'"' ); |
1303 | let suffix = s[s.len() - v.len() + 1..].to_owned().into_boxed_str(); |
1304 | (out, suffix) |
1305 | } |
1306 | |
1307 | fn parse_lit_byte_str_raw(s: &str) -> (Vec<u8>, Box<str>) { |
1308 | assert_eq!(byte(s, 0), b'b' ); |
1309 | let (value, suffix) = parse_lit_str_raw(&s[1..]); |
1310 | (String::from(value).into_bytes(), suffix) |
1311 | } |
1312 | |
1313 | // Returns (value, suffix). |
1314 | pub(crate) fn parse_lit_byte(s: &str) -> (u8, Box<str>) { |
1315 | assert_eq!(byte(s, 0), b'b' ); |
1316 | assert_eq!(byte(s, 1), b' \'' ); |
1317 | |
1318 | // We're going to want to have slices which don't respect codepoint boundaries. |
1319 | let mut v = s[2..].as_bytes(); |
1320 | |
1321 | let b = match byte(v, 0) { |
1322 | b' \\' => { |
1323 | let b = byte(v, 1); |
1324 | v = &v[2..]; |
1325 | match b { |
1326 | b'x' => { |
1327 | let (b, rest) = backslash_x(v); |
1328 | v = rest; |
1329 | b |
1330 | } |
1331 | b'n' => b' \n' , |
1332 | b'r' => b' \r' , |
1333 | b't' => b' \t' , |
1334 | b' \\' => b' \\' , |
1335 | b'0' => b' \0' , |
1336 | b' \'' => b' \'' , |
1337 | b'"' => b'"' , |
1338 | b => panic!("unexpected byte {:?} after \\ character in byte literal" , b), |
1339 | } |
1340 | } |
1341 | b => { |
1342 | v = &v[1..]; |
1343 | b |
1344 | } |
1345 | }; |
1346 | |
1347 | assert_eq!(byte(v, 0), b' \'' ); |
1348 | let suffix = s[s.len() - v.len() + 1..].to_owned().into_boxed_str(); |
1349 | (b, suffix) |
1350 | } |
1351 | |
1352 | // Returns (value, suffix). |
1353 | pub(crate) fn parse_lit_char(mut s: &str) -> (char, Box<str>) { |
1354 | assert_eq!(byte(s, 0), b' \'' ); |
1355 | s = &s[1..]; |
1356 | |
1357 | let ch = match byte(s, 0) { |
1358 | b' \\' => { |
1359 | let b = byte(s, 1); |
1360 | s = &s[2..]; |
1361 | match b { |
1362 | b'x' => { |
1363 | let (byte, rest) = backslash_x(s); |
1364 | s = rest; |
1365 | assert!(byte <= 0x80, "Invalid \\x byte in string literal" ); |
1366 | char::from_u32(u32::from(byte)).unwrap() |
1367 | } |
1368 | b'u' => { |
1369 | let (chr, rest) = backslash_u(s); |
1370 | s = rest; |
1371 | chr |
1372 | } |
1373 | b'n' => ' \n' , |
1374 | b'r' => ' \r' , |
1375 | b't' => ' \t' , |
1376 | b' \\' => ' \\' , |
1377 | b'0' => ' \0' , |
1378 | b' \'' => ' \'' , |
1379 | b'"' => '"' , |
1380 | b => panic!("unexpected byte {:?} after \\ character in byte literal" , b), |
1381 | } |
1382 | } |
1383 | _ => { |
1384 | let ch = next_chr(s); |
1385 | s = &s[ch.len_utf8()..]; |
1386 | ch |
1387 | } |
1388 | }; |
1389 | assert_eq!(byte(s, 0), b' \'' ); |
1390 | let suffix = s[1..].to_owned().into_boxed_str(); |
1391 | (ch, suffix) |
1392 | } |
1393 | |
1394 | fn backslash_x<S>(s: &S) -> (u8, &S) |
1395 | where |
1396 | S: Index<RangeFrom<usize>, Output = S> + AsRef<[u8]> + ?Sized, |
1397 | { |
1398 | let mut ch = 0; |
1399 | let b0 = byte(s, 0); |
1400 | let b1 = byte(s, 1); |
1401 | ch += 0x10 |
1402 | * match b0 { |
1403 | b'0' ..=b'9' => b0 - b'0' , |
1404 | b'a' ..=b'f' => 10 + (b0 - b'a' ), |
1405 | b'A' ..=b'F' => 10 + (b0 - b'A' ), |
1406 | _ => panic!("unexpected non-hex character after \\x" ), |
1407 | }; |
1408 | ch += match b1 { |
1409 | b'0' ..=b'9' => b1 - b'0' , |
1410 | b'a' ..=b'f' => 10 + (b1 - b'a' ), |
1411 | b'A' ..=b'F' => 10 + (b1 - b'A' ), |
1412 | _ => panic!("unexpected non-hex character after \\x" ), |
1413 | }; |
1414 | (ch, &s[2..]) |
1415 | } |
1416 | |
1417 | fn backslash_u(mut s: &str) -> (char, &str) { |
1418 | if byte(s, 0) != b'{' { |
1419 | panic!("{}" , "expected { after \\u" ); |
1420 | } |
1421 | s = &s[1..]; |
1422 | |
1423 | let mut ch = 0; |
1424 | let mut digits = 0; |
1425 | loop { |
1426 | let b = byte(s, 0); |
1427 | let digit = match b { |
1428 | b'0' ..=b'9' => b - b'0' , |
1429 | b'a' ..=b'f' => 10 + b - b'a' , |
1430 | b'A' ..=b'F' => 10 + b - b'A' , |
1431 | b'_' if digits > 0 => { |
1432 | s = &s[1..]; |
1433 | continue; |
1434 | } |
1435 | b'}' if digits == 0 => panic!("invalid empty unicode escape" ), |
1436 | b'}' => break, |
1437 | _ => panic!("unexpected non-hex character after \\u" ), |
1438 | }; |
1439 | if digits == 6 { |
1440 | panic!("overlong unicode escape (must have at most 6 hex digits)" ); |
1441 | } |
1442 | ch *= 0x10; |
1443 | ch += u32::from(digit); |
1444 | digits += 1; |
1445 | s = &s[1..]; |
1446 | } |
1447 | assert!(byte(s, 0) == b'}' ); |
1448 | s = &s[1..]; |
1449 | |
1450 | if let Some(ch) = char::from_u32(ch) { |
1451 | (ch, s) |
1452 | } else { |
1453 | panic!("character code {:x} is not a valid unicode character" , ch); |
1454 | } |
1455 | } |
1456 | |
1457 | // Returns base 10 digits and suffix. |
1458 | pub(crate) fn parse_lit_int(mut s: &str) -> Option<(Box<str>, Box<str>)> { |
1459 | let negative = byte(s, 0) == b'-' ; |
1460 | if negative { |
1461 | s = &s[1..]; |
1462 | } |
1463 | |
1464 | let base = match (byte(s, 0), byte(s, 1)) { |
1465 | (b'0' , b'x' ) => { |
1466 | s = &s[2..]; |
1467 | 16 |
1468 | } |
1469 | (b'0' , b'o' ) => { |
1470 | s = &s[2..]; |
1471 | 8 |
1472 | } |
1473 | (b'0' , b'b' ) => { |
1474 | s = &s[2..]; |
1475 | 2 |
1476 | } |
1477 | (b'0' ..=b'9' , _) => 10, |
1478 | _ => return None, |
1479 | }; |
1480 | |
1481 | let mut value = BigInt::new(); |
1482 | let mut has_digit = false; |
1483 | 'outer: loop { |
1484 | let b = byte(s, 0); |
1485 | let digit = match b { |
1486 | b'0' ..=b'9' => b - b'0' , |
1487 | b'a' ..=b'f' if base > 10 => b - b'a' + 10, |
1488 | b'A' ..=b'F' if base > 10 => b - b'A' + 10, |
1489 | b'_' => { |
1490 | s = &s[1..]; |
1491 | continue; |
1492 | } |
1493 | // If looking at a floating point literal, we don't want to |
1494 | // consider it an integer. |
1495 | b'.' if base == 10 => return None, |
1496 | b'e' | b'E' if base == 10 => { |
1497 | let mut has_exp = false; |
1498 | for (i, b) in s[1..].bytes().enumerate() { |
1499 | match b { |
1500 | b'_' => {} |
1501 | b'-' | b'+' => return None, |
1502 | b'0' ..=b'9' => has_exp = true, |
1503 | _ => { |
1504 | let suffix = &s[1 + i..]; |
1505 | if has_exp && crate::ident::xid_ok(suffix) { |
1506 | return None; |
1507 | } else { |
1508 | break 'outer; |
1509 | } |
1510 | } |
1511 | } |
1512 | } |
1513 | if has_exp { |
1514 | return None; |
1515 | } else { |
1516 | break; |
1517 | } |
1518 | } |
1519 | _ => break, |
1520 | }; |
1521 | |
1522 | if digit >= base { |
1523 | return None; |
1524 | } |
1525 | |
1526 | has_digit = true; |
1527 | value *= base; |
1528 | value += digit; |
1529 | s = &s[1..]; |
1530 | } |
1531 | |
1532 | if !has_digit { |
1533 | return None; |
1534 | } |
1535 | |
1536 | let suffix = s; |
1537 | if suffix.is_empty() || crate::ident::xid_ok(suffix) { |
1538 | let mut repr = value.to_string(); |
1539 | if negative { |
1540 | repr.insert(0, '-' ); |
1541 | } |
1542 | Some((repr.into_boxed_str(), suffix.to_owned().into_boxed_str())) |
1543 | } else { |
1544 | None |
1545 | } |
1546 | } |
1547 | |
1548 | // Returns base 10 digits and suffix. |
1549 | pub(crate) fn parse_lit_float(input: &str) -> Option<(Box<str>, Box<str>)> { |
1550 | // Rust's floating point literals are very similar to the ones parsed by |
1551 | // the standard library, except that rust's literals can contain |
1552 | // ignorable underscores. Let's remove those underscores. |
1553 | |
1554 | let mut bytes = input.to_owned().into_bytes(); |
1555 | |
1556 | let start = (*bytes.first()? == b'-' ) as usize; |
1557 | match bytes.get(start)? { |
1558 | b'0' ..=b'9' => {} |
1559 | _ => return None, |
1560 | } |
1561 | |
1562 | let mut read = start; |
1563 | let mut write = start; |
1564 | let mut has_dot = false; |
1565 | let mut has_e = false; |
1566 | let mut has_sign = false; |
1567 | let mut has_exponent = false; |
1568 | while read < bytes.len() { |
1569 | match bytes[read] { |
1570 | b'_' => { |
1571 | // Don't increase write |
1572 | read += 1; |
1573 | continue; |
1574 | } |
1575 | b'0' ..=b'9' => { |
1576 | if has_e { |
1577 | has_exponent = true; |
1578 | } |
1579 | bytes[write] = bytes[read]; |
1580 | } |
1581 | b'.' => { |
1582 | if has_e || has_dot { |
1583 | return None; |
1584 | } |
1585 | has_dot = true; |
1586 | bytes[write] = b'.' ; |
1587 | } |
1588 | b'e' | b'E' => { |
1589 | match bytes[read + 1..] |
1590 | .iter() |
1591 | .find(|b| **b != b'_' ) |
1592 | .unwrap_or(&b' \0' ) |
1593 | { |
1594 | b'-' | b'+' | b'0' ..=b'9' => {} |
1595 | _ => break, |
1596 | } |
1597 | if has_e { |
1598 | if has_exponent { |
1599 | break; |
1600 | } else { |
1601 | return None; |
1602 | } |
1603 | } |
1604 | has_e = true; |
1605 | bytes[write] = b'e' ; |
1606 | } |
1607 | b'-' | b'+' => { |
1608 | if has_sign || has_exponent || !has_e { |
1609 | return None; |
1610 | } |
1611 | has_sign = true; |
1612 | if bytes[read] == b'-' { |
1613 | bytes[write] = bytes[read]; |
1614 | } else { |
1615 | // Omit '+' |
1616 | read += 1; |
1617 | continue; |
1618 | } |
1619 | } |
1620 | _ => break, |
1621 | } |
1622 | read += 1; |
1623 | write += 1; |
1624 | } |
1625 | |
1626 | if has_e && !has_exponent { |
1627 | return None; |
1628 | } |
1629 | |
1630 | let mut digits = String::from_utf8(bytes).unwrap(); |
1631 | let suffix = digits.split_off(read); |
1632 | digits.truncate(write); |
1633 | if suffix.is_empty() || crate::ident::xid_ok(&suffix) { |
1634 | Some((digits.into_boxed_str(), suffix.into_boxed_str())) |
1635 | } else { |
1636 | None |
1637 | } |
1638 | } |
1639 | } |
1640 | |