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::ffi::{CStr, CString}; |
10 | use std::fmt::{self, Display}; |
11 | #[cfg (feature = "extra-traits" )] |
12 | use std::hash::{Hash, Hasher}; |
13 | use std::str::{self, FromStr}; |
14 | |
15 | ast_enum_of_structs! { |
16 | /// A Rust literal such as a string or integer or boolean. |
17 | /// |
18 | /// # Syntax tree enum |
19 | /// |
20 | /// This type is a [syntax tree enum]. |
21 | /// |
22 | /// [syntax tree enum]: crate::expr::Expr#syntax-tree-enums |
23 | #[non_exhaustive ] |
24 | pub enum Lit { |
25 | /// A UTF-8 string literal: `"foo"`. |
26 | Str(LitStr), |
27 | |
28 | /// A byte string literal: `b"foo"`. |
29 | ByteStr(LitByteStr), |
30 | |
31 | /// A nul-terminated C-string literal: `c"foo"`. |
32 | CStr(LitCStr), |
33 | |
34 | /// A byte literal: `b'f'`. |
35 | Byte(LitByte), |
36 | |
37 | /// A character literal: `'a'`. |
38 | Char(LitChar), |
39 | |
40 | /// An integer literal: `1` or `1u16`. |
41 | Int(LitInt), |
42 | |
43 | /// A floating point literal: `1f64` or `1.0e10f64`. |
44 | /// |
45 | /// Must be finite. May not be infinite or NaN. |
46 | Float(LitFloat), |
47 | |
48 | /// A boolean literal: `true` or `false`. |
49 | Bool(LitBool), |
50 | |
51 | /// A raw token literal not interpreted by Syn. |
52 | Verbatim(Literal), |
53 | } |
54 | } |
55 | |
56 | ast_struct! { |
57 | /// A UTF-8 string literal: `"foo"`. |
58 | pub struct LitStr { |
59 | repr: Box<LitRepr>, |
60 | } |
61 | } |
62 | |
63 | ast_struct! { |
64 | /// A byte string literal: `b"foo"`. |
65 | pub struct LitByteStr { |
66 | repr: Box<LitRepr>, |
67 | } |
68 | } |
69 | |
70 | ast_struct! { |
71 | /// A nul-terminated C-string literal: `c"foo"`. |
72 | pub struct LitCStr { |
73 | repr: Box<LitRepr>, |
74 | } |
75 | } |
76 | |
77 | ast_struct! { |
78 | /// A byte literal: `b'f'`. |
79 | pub struct LitByte { |
80 | repr: Box<LitRepr>, |
81 | } |
82 | } |
83 | |
84 | ast_struct! { |
85 | /// A character literal: `'a'`. |
86 | pub struct LitChar { |
87 | repr: Box<LitRepr>, |
88 | } |
89 | } |
90 | |
91 | struct LitRepr { |
92 | token: Literal, |
93 | suffix: Box<str>, |
94 | } |
95 | |
96 | ast_struct! { |
97 | /// An integer literal: `1` or `1u16`. |
98 | pub struct LitInt { |
99 | repr: Box<LitIntRepr>, |
100 | } |
101 | } |
102 | |
103 | struct LitIntRepr { |
104 | token: Literal, |
105 | digits: Box<str>, |
106 | suffix: Box<str>, |
107 | } |
108 | |
109 | ast_struct! { |
110 | /// A floating point literal: `1f64` or `1.0e10f64`. |
111 | /// |
112 | /// Must be finite. May not be infinite or NaN. |
113 | pub struct LitFloat { |
114 | repr: Box<LitFloatRepr>, |
115 | } |
116 | } |
117 | |
118 | struct LitFloatRepr { |
119 | token: Literal, |
120 | digits: Box<str>, |
121 | suffix: Box<str>, |
122 | } |
123 | |
124 | ast_struct! { |
125 | /// A boolean literal: `true` or `false`. |
126 | pub struct LitBool { |
127 | pub value: bool, |
128 | pub span: Span, |
129 | } |
130 | } |
131 | |
132 | impl LitStr { |
133 | pub fn new(value: &str, span: Span) -> Self { |
134 | let mut token = Literal::string(value); |
135 | token.set_span(span); |
136 | LitStr { |
137 | repr: Box::new(LitRepr { |
138 | token, |
139 | suffix: Box::<str>::default(), |
140 | }), |
141 | } |
142 | } |
143 | |
144 | pub fn value(&self) -> String { |
145 | let repr = self.repr.token.to_string(); |
146 | let (value, _suffix) = value::parse_lit_str(&repr); |
147 | String::from(value) |
148 | } |
149 | |
150 | /// Parse a syntax tree node from the content of this string literal. |
151 | /// |
152 | /// All spans in the syntax tree will point to the span of this `LitStr`. |
153 | /// |
154 | /// # Example |
155 | /// |
156 | /// ``` |
157 | /// use syn::{Attribute, Error, Expr, Lit, Meta, Path, Result}; |
158 | /// |
159 | /// // Parses the path from an attribute that looks like: |
160 | /// // |
161 | /// // #[path = "a::b::c"] |
162 | /// // |
163 | /// // or returns `None` if the input is some other attribute. |
164 | /// fn get_path(attr: &Attribute) -> Result<Option<Path>> { |
165 | /// if !attr.path().is_ident("path" ) { |
166 | /// return Ok(None); |
167 | /// } |
168 | /// |
169 | /// if let Meta::NameValue(meta) = &attr.meta { |
170 | /// if let Expr::Lit(expr) = &meta.value { |
171 | /// if let Lit::Str(lit_str) = &expr.lit { |
172 | /// return lit_str.parse().map(Some); |
173 | /// } |
174 | /// } |
175 | /// } |
176 | /// |
177 | /// let message = "expected #[path = \"... \"]" ; |
178 | /// Err(Error::new_spanned(attr, message)) |
179 | /// } |
180 | /// ``` |
181 | #[cfg (feature = "parsing" )] |
182 | #[cfg_attr (docsrs, doc(cfg(feature = "parsing" )))] |
183 | pub fn parse<T: Parse>(&self) -> Result<T> { |
184 | self.parse_with(T::parse) |
185 | } |
186 | |
187 | /// Invoke parser on the content of this string literal. |
188 | /// |
189 | /// All spans in the syntax tree will point to the span of this `LitStr`. |
190 | /// |
191 | /// # Example |
192 | /// |
193 | /// ``` |
194 | /// # use proc_macro2::Span; |
195 | /// # use syn::{LitStr, Result}; |
196 | /// # |
197 | /// # fn main() -> Result<()> { |
198 | /// # let lit_str = LitStr::new("a::b::c" , Span::call_site()); |
199 | /// # |
200 | /// # const IGNORE: &str = stringify! { |
201 | /// let lit_str: LitStr = /* ... */; |
202 | /// # }; |
203 | /// |
204 | /// // Parse a string literal like "a::b::c" into a Path, not allowing |
205 | /// // generic arguments on any of the path segments. |
206 | /// let basic_path = lit_str.parse_with(syn::Path::parse_mod_style)?; |
207 | /// # |
208 | /// # Ok(()) |
209 | /// # } |
210 | /// ``` |
211 | #[cfg (feature = "parsing" )] |
212 | #[cfg_attr (docsrs, doc(cfg(feature = "parsing" )))] |
213 | pub fn parse_with<F: Parser>(&self, parser: F) -> Result<F::Output> { |
214 | use proc_macro2::Group; |
215 | |
216 | // Token stream with every span replaced by the given one. |
217 | fn respan_token_stream(stream: TokenStream, span: Span) -> TokenStream { |
218 | stream |
219 | .into_iter() |
220 | .map(|token| respan_token_tree(token, span)) |
221 | .collect() |
222 | } |
223 | |
224 | // Token tree with every span replaced by the given one. |
225 | fn respan_token_tree(mut token: TokenTree, span: Span) -> TokenTree { |
226 | match &mut token { |
227 | TokenTree::Group(g) => { |
228 | let stream = respan_token_stream(g.stream(), span); |
229 | *g = Group::new(g.delimiter(), stream); |
230 | g.set_span(span); |
231 | } |
232 | other => other.set_span(span), |
233 | } |
234 | token |
235 | } |
236 | |
237 | // Parse string literal into a token stream with every span equal to the |
238 | // original literal's span. |
239 | let span = self.span(); |
240 | let mut tokens = TokenStream::from_str(&self.value())?; |
241 | tokens = respan_token_stream(tokens, span); |
242 | |
243 | let result = crate::parse::parse_scoped(parser, span, tokens)?; |
244 | |
245 | let suffix = self.suffix(); |
246 | if !suffix.is_empty() { |
247 | return Err(Error::new( |
248 | self.span(), |
249 | format!("unexpected suffix ` {}` on string literal" , suffix), |
250 | )); |
251 | } |
252 | |
253 | Ok(result) |
254 | } |
255 | |
256 | pub fn span(&self) -> Span { |
257 | self.repr.token.span() |
258 | } |
259 | |
260 | pub fn set_span(&mut self, span: Span) { |
261 | self.repr.token.set_span(span); |
262 | } |
263 | |
264 | pub fn suffix(&self) -> &str { |
265 | &self.repr.suffix |
266 | } |
267 | |
268 | pub fn token(&self) -> Literal { |
269 | self.repr.token.clone() |
270 | } |
271 | } |
272 | |
273 | impl LitByteStr { |
274 | pub fn new(value: &[u8], span: Span) -> Self { |
275 | let mut token = Literal::byte_string(value); |
276 | token.set_span(span); |
277 | LitByteStr { |
278 | repr: Box::new(LitRepr { |
279 | token, |
280 | suffix: Box::<str>::default(), |
281 | }), |
282 | } |
283 | } |
284 | |
285 | pub fn value(&self) -> Vec<u8> { |
286 | let repr = self.repr.token.to_string(); |
287 | let (value, _suffix) = value::parse_lit_byte_str(&repr); |
288 | value |
289 | } |
290 | |
291 | pub fn span(&self) -> Span { |
292 | self.repr.token.span() |
293 | } |
294 | |
295 | pub fn set_span(&mut self, span: Span) { |
296 | self.repr.token.set_span(span); |
297 | } |
298 | |
299 | pub fn suffix(&self) -> &str { |
300 | &self.repr.suffix |
301 | } |
302 | |
303 | pub fn token(&self) -> Literal { |
304 | self.repr.token.clone() |
305 | } |
306 | } |
307 | |
308 | impl LitCStr { |
309 | pub fn new(value: &CStr, span: Span) -> Self { |
310 | let mut token = Literal::c_string(value); |
311 | token.set_span(span); |
312 | LitCStr { |
313 | repr: Box::new(LitRepr { |
314 | token, |
315 | suffix: Box::<str>::default(), |
316 | }), |
317 | } |
318 | } |
319 | |
320 | pub fn value(&self) -> CString { |
321 | let repr = self.repr.token.to_string(); |
322 | let (value, _suffix) = value::parse_lit_c_str(&repr); |
323 | value |
324 | } |
325 | |
326 | pub fn span(&self) -> Span { |
327 | self.repr.token.span() |
328 | } |
329 | |
330 | pub fn set_span(&mut self, span: Span) { |
331 | self.repr.token.set_span(span); |
332 | } |
333 | |
334 | pub fn suffix(&self) -> &str { |
335 | &self.repr.suffix |
336 | } |
337 | |
338 | pub fn token(&self) -> Literal { |
339 | self.repr.token.clone() |
340 | } |
341 | } |
342 | |
343 | impl LitByte { |
344 | pub fn new(value: u8, span: Span) -> Self { |
345 | let mut token = Literal::u8_suffixed(value); |
346 | token.set_span(span); |
347 | LitByte { |
348 | repr: Box::new(LitRepr { |
349 | token, |
350 | suffix: Box::<str>::default(), |
351 | }), |
352 | } |
353 | } |
354 | |
355 | pub fn value(&self) -> u8 { |
356 | let repr = self.repr.token.to_string(); |
357 | let (value, _suffix) = value::parse_lit_byte(&repr); |
358 | value |
359 | } |
360 | |
361 | pub fn span(&self) -> Span { |
362 | self.repr.token.span() |
363 | } |
364 | |
365 | pub fn set_span(&mut self, span: Span) { |
366 | self.repr.token.set_span(span); |
367 | } |
368 | |
369 | pub fn suffix(&self) -> &str { |
370 | &self.repr.suffix |
371 | } |
372 | |
373 | pub fn token(&self) -> Literal { |
374 | self.repr.token.clone() |
375 | } |
376 | } |
377 | |
378 | impl LitChar { |
379 | pub fn new(value: char, span: Span) -> Self { |
380 | let mut token = Literal::character(value); |
381 | token.set_span(span); |
382 | LitChar { |
383 | repr: Box::new(LitRepr { |
384 | token, |
385 | suffix: Box::<str>::default(), |
386 | }), |
387 | } |
388 | } |
389 | |
390 | pub fn value(&self) -> char { |
391 | let repr = self.repr.token.to_string(); |
392 | let (value, _suffix) = value::parse_lit_char(&repr); |
393 | value |
394 | } |
395 | |
396 | pub fn span(&self) -> Span { |
397 | self.repr.token.span() |
398 | } |
399 | |
400 | pub fn set_span(&mut self, span: Span) { |
401 | self.repr.token.set_span(span); |
402 | } |
403 | |
404 | pub fn suffix(&self) -> &str { |
405 | &self.repr.suffix |
406 | } |
407 | |
408 | pub fn token(&self) -> Literal { |
409 | self.repr.token.clone() |
410 | } |
411 | } |
412 | |
413 | impl LitInt { |
414 | pub fn new(repr: &str, span: Span) -> Self { |
415 | let (digits, suffix) = match value::parse_lit_int(repr) { |
416 | Some(parse) => parse, |
417 | None => panic!("not an integer literal: ` {}`" , repr), |
418 | }; |
419 | |
420 | let mut token: Literal = repr.parse().unwrap(); |
421 | token.set_span(span); |
422 | LitInt { |
423 | repr: Box::new(LitIntRepr { |
424 | token, |
425 | digits, |
426 | suffix, |
427 | }), |
428 | } |
429 | } |
430 | |
431 | pub fn base10_digits(&self) -> &str { |
432 | &self.repr.digits |
433 | } |
434 | |
435 | /// Parses the literal into a selected number type. |
436 | /// |
437 | /// This is equivalent to `lit.base10_digits().parse()` except that the |
438 | /// resulting errors will be correctly spanned to point to the literal token |
439 | /// in the macro input. |
440 | /// |
441 | /// ``` |
442 | /// use syn::LitInt; |
443 | /// use syn::parse::{Parse, ParseStream, Result}; |
444 | /// |
445 | /// struct Port { |
446 | /// value: u16, |
447 | /// } |
448 | /// |
449 | /// impl Parse for Port { |
450 | /// fn parse(input: ParseStream) -> Result<Self> { |
451 | /// let lit: LitInt = input.parse()?; |
452 | /// let value = lit.base10_parse::<u16>()?; |
453 | /// Ok(Port { value }) |
454 | /// } |
455 | /// } |
456 | /// ``` |
457 | pub fn base10_parse<N>(&self) -> Result<N> |
458 | where |
459 | N: FromStr, |
460 | N::Err: Display, |
461 | { |
462 | self.base10_digits() |
463 | .parse() |
464 | .map_err(|err| Error::new(self.span(), err)) |
465 | } |
466 | |
467 | pub fn suffix(&self) -> &str { |
468 | &self.repr.suffix |
469 | } |
470 | |
471 | pub fn span(&self) -> Span { |
472 | self.repr.token.span() |
473 | } |
474 | |
475 | pub fn set_span(&mut self, span: Span) { |
476 | self.repr.token.set_span(span); |
477 | } |
478 | |
479 | pub fn token(&self) -> Literal { |
480 | self.repr.token.clone() |
481 | } |
482 | } |
483 | |
484 | impl From<Literal> for LitInt { |
485 | fn from(token: Literal) -> Self { |
486 | let repr: String = token.to_string(); |
487 | if let Some((digits: Box, suffix: Box)) = value::parse_lit_int(&repr) { |
488 | LitInt { |
489 | repr: Box::new(LitIntRepr { |
490 | token, |
491 | digits, |
492 | suffix, |
493 | }), |
494 | } |
495 | } else { |
496 | panic!("not an integer literal: ` {}`" , repr); |
497 | } |
498 | } |
499 | } |
500 | |
501 | impl Display for LitInt { |
502 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
503 | self.repr.token.fmt(formatter) |
504 | } |
505 | } |
506 | |
507 | impl LitFloat { |
508 | pub fn new(repr: &str, span: Span) -> Self { |
509 | let (digits, suffix) = match value::parse_lit_float(repr) { |
510 | Some(parse) => parse, |
511 | None => panic!("not a float literal: ` {}`" , repr), |
512 | }; |
513 | |
514 | let mut token: Literal = repr.parse().unwrap(); |
515 | token.set_span(span); |
516 | LitFloat { |
517 | repr: Box::new(LitFloatRepr { |
518 | token, |
519 | digits, |
520 | suffix, |
521 | }), |
522 | } |
523 | } |
524 | |
525 | pub fn base10_digits(&self) -> &str { |
526 | &self.repr.digits |
527 | } |
528 | |
529 | pub fn base10_parse<N>(&self) -> Result<N> |
530 | where |
531 | N: FromStr, |
532 | N::Err: Display, |
533 | { |
534 | self.base10_digits() |
535 | .parse() |
536 | .map_err(|err| Error::new(self.span(), err)) |
537 | } |
538 | |
539 | pub fn suffix(&self) -> &str { |
540 | &self.repr.suffix |
541 | } |
542 | |
543 | pub fn span(&self) -> Span { |
544 | self.repr.token.span() |
545 | } |
546 | |
547 | pub fn set_span(&mut self, span: Span) { |
548 | self.repr.token.set_span(span); |
549 | } |
550 | |
551 | pub fn token(&self) -> Literal { |
552 | self.repr.token.clone() |
553 | } |
554 | } |
555 | |
556 | impl From<Literal> for LitFloat { |
557 | fn from(token: Literal) -> Self { |
558 | let repr: String = token.to_string(); |
559 | if let Some((digits: Box, suffix: Box)) = value::parse_lit_float(&repr) { |
560 | LitFloat { |
561 | repr: Box::new(LitFloatRepr { |
562 | token, |
563 | digits, |
564 | suffix, |
565 | }), |
566 | } |
567 | } else { |
568 | panic!("not a float literal: ` {}`" , repr); |
569 | } |
570 | } |
571 | } |
572 | |
573 | impl Display for LitFloat { |
574 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
575 | self.repr.token.fmt(formatter) |
576 | } |
577 | } |
578 | |
579 | impl LitBool { |
580 | pub fn new(value: bool, span: Span) -> Self { |
581 | LitBool { value, span } |
582 | } |
583 | |
584 | pub fn value(&self) -> bool { |
585 | self.value |
586 | } |
587 | |
588 | pub fn span(&self) -> Span { |
589 | self.span |
590 | } |
591 | |
592 | pub fn set_span(&mut self, span: Span) { |
593 | self.span = span; |
594 | } |
595 | |
596 | pub fn token(&self) -> Ident { |
597 | let s: &'static str = if self.value { "true" } else { "false" }; |
598 | Ident::new(string:s, self.span) |
599 | } |
600 | } |
601 | |
602 | #[cfg (feature = "extra-traits" )] |
603 | mod debug_impls { |
604 | use crate::lit::{LitBool, LitByte, LitByteStr, LitCStr, LitChar, LitFloat, LitInt, LitStr}; |
605 | use std::fmt::{self, Debug}; |
606 | |
607 | #[cfg_attr (docsrs, doc(cfg(feature = "extra-traits" )))] |
608 | impl Debug for LitStr { |
609 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
610 | self.debug(formatter, "LitStr" ) |
611 | } |
612 | } |
613 | |
614 | impl LitStr { |
615 | pub(crate) fn debug(&self, formatter: &mut fmt::Formatter, name: &str) -> fmt::Result { |
616 | formatter |
617 | .debug_struct(name) |
618 | .field("token" , &format_args!(" {}" , self.repr.token)) |
619 | .finish() |
620 | } |
621 | } |
622 | |
623 | #[cfg_attr (docsrs, doc(cfg(feature = "extra-traits" )))] |
624 | impl Debug for LitByteStr { |
625 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
626 | self.debug(formatter, "LitByteStr" ) |
627 | } |
628 | } |
629 | |
630 | impl LitByteStr { |
631 | pub(crate) fn debug(&self, formatter: &mut fmt::Formatter, name: &str) -> fmt::Result { |
632 | formatter |
633 | .debug_struct(name) |
634 | .field("token" , &format_args!(" {}" , self.repr.token)) |
635 | .finish() |
636 | } |
637 | } |
638 | |
639 | #[cfg_attr (docsrs, doc(cfg(feature = "extra-traits" )))] |
640 | impl Debug for LitCStr { |
641 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
642 | self.debug(formatter, "LitCStr" ) |
643 | } |
644 | } |
645 | |
646 | impl LitCStr { |
647 | pub(crate) fn debug(&self, formatter: &mut fmt::Formatter, name: &str) -> fmt::Result { |
648 | formatter |
649 | .debug_struct(name) |
650 | .field("token" , &format_args!(" {}" , self.repr.token)) |
651 | .finish() |
652 | } |
653 | } |
654 | |
655 | #[cfg_attr (docsrs, doc(cfg(feature = "extra-traits" )))] |
656 | impl Debug for LitByte { |
657 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
658 | self.debug(formatter, "LitByte" ) |
659 | } |
660 | } |
661 | |
662 | impl LitByte { |
663 | pub(crate) fn debug(&self, formatter: &mut fmt::Formatter, name: &str) -> fmt::Result { |
664 | formatter |
665 | .debug_struct(name) |
666 | .field("token" , &format_args!(" {}" , self.repr.token)) |
667 | .finish() |
668 | } |
669 | } |
670 | |
671 | #[cfg_attr (docsrs, doc(cfg(feature = "extra-traits" )))] |
672 | impl Debug for LitChar { |
673 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
674 | self.debug(formatter, "LitChar" ) |
675 | } |
676 | } |
677 | |
678 | impl LitChar { |
679 | pub(crate) fn debug(&self, formatter: &mut fmt::Formatter, name: &str) -> fmt::Result { |
680 | formatter |
681 | .debug_struct(name) |
682 | .field("token" , &format_args!(" {}" , self.repr.token)) |
683 | .finish() |
684 | } |
685 | } |
686 | |
687 | #[cfg_attr (docsrs, doc(cfg(feature = "extra-traits" )))] |
688 | impl Debug for LitInt { |
689 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
690 | self.debug(formatter, "LitInt" ) |
691 | } |
692 | } |
693 | |
694 | impl LitInt { |
695 | pub(crate) fn debug(&self, formatter: &mut fmt::Formatter, name: &str) -> fmt::Result { |
696 | formatter |
697 | .debug_struct(name) |
698 | .field("token" , &format_args!(" {}" , self.repr.token)) |
699 | .finish() |
700 | } |
701 | } |
702 | |
703 | #[cfg_attr (docsrs, doc(cfg(feature = "extra-traits" )))] |
704 | impl Debug for LitFloat { |
705 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
706 | self.debug(formatter, "LitFloat" ) |
707 | } |
708 | } |
709 | |
710 | impl LitFloat { |
711 | pub(crate) fn debug(&self, formatter: &mut fmt::Formatter, name: &str) -> fmt::Result { |
712 | formatter |
713 | .debug_struct(name) |
714 | .field("token" , &format_args!(" {}" , self.repr.token)) |
715 | .finish() |
716 | } |
717 | } |
718 | |
719 | #[cfg_attr (docsrs, doc(cfg(feature = "extra-traits" )))] |
720 | impl Debug for LitBool { |
721 | fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result { |
722 | self.debug(formatter, "LitBool" ) |
723 | } |
724 | } |
725 | |
726 | impl LitBool { |
727 | pub(crate) fn debug(&self, formatter: &mut fmt::Formatter, name: &str) -> fmt::Result { |
728 | formatter |
729 | .debug_struct(name) |
730 | .field("value" , &self.value) |
731 | .finish() |
732 | } |
733 | } |
734 | } |
735 | |
736 | #[cfg (feature = "clone-impls" )] |
737 | #[cfg_attr (docsrs, doc(cfg(feature = "clone-impls" )))] |
738 | impl Clone for LitRepr { |
739 | fn clone(&self) -> Self { |
740 | LitRepr { |
741 | token: self.token.clone(), |
742 | suffix: self.suffix.clone(), |
743 | } |
744 | } |
745 | } |
746 | |
747 | #[cfg (feature = "clone-impls" )] |
748 | #[cfg_attr (docsrs, doc(cfg(feature = "clone-impls" )))] |
749 | impl Clone for LitIntRepr { |
750 | fn clone(&self) -> Self { |
751 | LitIntRepr { |
752 | token: self.token.clone(), |
753 | digits: self.digits.clone(), |
754 | suffix: self.suffix.clone(), |
755 | } |
756 | } |
757 | } |
758 | |
759 | #[cfg (feature = "clone-impls" )] |
760 | #[cfg_attr (docsrs, doc(cfg(feature = "clone-impls" )))] |
761 | impl Clone for LitFloatRepr { |
762 | fn clone(&self) -> Self { |
763 | LitFloatRepr { |
764 | token: self.token.clone(), |
765 | digits: self.digits.clone(), |
766 | suffix: self.suffix.clone(), |
767 | } |
768 | } |
769 | } |
770 | |
771 | macro_rules! lit_extra_traits { |
772 | ($ty:ident) => { |
773 | #[cfg(feature = "clone-impls" )] |
774 | #[cfg_attr(docsrs, doc(cfg(feature = "clone-impls" )))] |
775 | impl Clone for $ty { |
776 | fn clone(&self) -> Self { |
777 | $ty { |
778 | repr: self.repr.clone(), |
779 | } |
780 | } |
781 | } |
782 | |
783 | #[cfg(feature = "extra-traits" )] |
784 | #[cfg_attr(docsrs, doc(cfg(feature = "extra-traits" )))] |
785 | impl PartialEq for $ty { |
786 | fn eq(&self, other: &Self) -> bool { |
787 | self.repr.token.to_string() == other.repr.token.to_string() |
788 | } |
789 | } |
790 | |
791 | #[cfg(feature = "extra-traits" )] |
792 | #[cfg_attr(docsrs, doc(cfg(feature = "extra-traits" )))] |
793 | impl Hash for $ty { |
794 | fn hash<H>(&self, state: &mut H) |
795 | where |
796 | H: Hasher, |
797 | { |
798 | self.repr.token.to_string().hash(state); |
799 | } |
800 | } |
801 | |
802 | #[cfg(feature = "parsing" )] |
803 | pub_if_not_doc! { |
804 | #[doc(hidden)] |
805 | #[allow(non_snake_case)] |
806 | pub fn $ty(marker: lookahead::TokenMarker) -> $ty { |
807 | match marker {} |
808 | } |
809 | } |
810 | }; |
811 | } |
812 | |
813 | lit_extra_traits!(LitStr); |
814 | lit_extra_traits!(LitByteStr); |
815 | lit_extra_traits!(LitCStr); |
816 | lit_extra_traits!(LitByte); |
817 | lit_extra_traits!(LitChar); |
818 | lit_extra_traits!(LitInt); |
819 | lit_extra_traits!(LitFloat); |
820 | |
821 | #[cfg (feature = "parsing" )] |
822 | pub_if_not_doc! { |
823 | #[doc (hidden)] |
824 | #[allow (non_snake_case)] |
825 | pub fn LitBool(marker: lookahead::TokenMarker) -> LitBool { |
826 | match marker {} |
827 | } |
828 | } |
829 | |
830 | /// The style of a string literal, either plain quoted or a raw string like |
831 | /// `r##"data"##`. |
832 | #[doc (hidden)] // https://github.com/dtolnay/syn/issues/1566 |
833 | pub enum StrStyle { |
834 | /// An ordinary string like `"data"`. |
835 | Cooked, |
836 | /// A raw string like `r##"data"##`. |
837 | /// |
838 | /// The unsigned integer is the number of `#` symbols used. |
839 | Raw(usize), |
840 | } |
841 | |
842 | #[cfg (feature = "parsing" )] |
843 | pub_if_not_doc! { |
844 | #[doc (hidden)] |
845 | #[allow (non_snake_case)] |
846 | pub fn Lit(marker: lookahead::TokenMarker) -> Lit { |
847 | match marker {} |
848 | } |
849 | } |
850 | |
851 | #[cfg (feature = "parsing" )] |
852 | pub(crate) mod parsing { |
853 | use crate::buffer::Cursor; |
854 | use crate::error::Result; |
855 | use crate::lit::{ |
856 | value, Lit, LitBool, LitByte, LitByteStr, LitCStr, LitChar, LitFloat, LitFloatRepr, LitInt, |
857 | LitIntRepr, LitStr, |
858 | }; |
859 | use crate::parse::{Parse, ParseStream, Unexpected}; |
860 | use crate::token::{self, Token}; |
861 | use proc_macro2::{Literal, Punct, Span}; |
862 | use std::cell::Cell; |
863 | use std::rc::Rc; |
864 | |
865 | #[cfg_attr (docsrs, doc(cfg(feature = "parsing" )))] |
866 | impl Parse for Lit { |
867 | fn parse(input: ParseStream) -> Result<Self> { |
868 | input.step(|cursor| { |
869 | if let Some((lit, rest)) = cursor.literal() { |
870 | return Ok((Lit::new(lit), rest)); |
871 | } |
872 | |
873 | if let Some((ident, rest)) = cursor.ident() { |
874 | let value = ident == "true" ; |
875 | if value || ident == "false" { |
876 | let lit_bool = LitBool { |
877 | value, |
878 | span: ident.span(), |
879 | }; |
880 | return Ok((Lit::Bool(lit_bool), rest)); |
881 | } |
882 | } |
883 | |
884 | if let Some((punct, rest)) = cursor.punct() { |
885 | if punct.as_char() == '-' { |
886 | if let Some((lit, rest)) = parse_negative_lit(punct, rest) { |
887 | return Ok((lit, rest)); |
888 | } |
889 | } |
890 | } |
891 | |
892 | Err(cursor.error("expected literal" )) |
893 | }) |
894 | } |
895 | } |
896 | |
897 | fn parse_negative_lit(neg: Punct, cursor: Cursor) -> Option<(Lit, Cursor)> { |
898 | let (lit, rest) = cursor.literal()?; |
899 | |
900 | let mut span = neg.span(); |
901 | span = span.join(lit.span()).unwrap_or(span); |
902 | |
903 | let mut repr = lit.to_string(); |
904 | repr.insert(0, '-' ); |
905 | |
906 | if let Some((digits, suffix)) = value::parse_lit_int(&repr) { |
907 | let mut token: Literal = repr.parse().unwrap(); |
908 | token.set_span(span); |
909 | return Some(( |
910 | Lit::Int(LitInt { |
911 | repr: Box::new(LitIntRepr { |
912 | token, |
913 | digits, |
914 | suffix, |
915 | }), |
916 | }), |
917 | rest, |
918 | )); |
919 | } |
920 | |
921 | let (digits, suffix) = value::parse_lit_float(&repr)?; |
922 | let mut token: Literal = repr.parse().unwrap(); |
923 | token.set_span(span); |
924 | Some(( |
925 | Lit::Float(LitFloat { |
926 | repr: Box::new(LitFloatRepr { |
927 | token, |
928 | digits, |
929 | suffix, |
930 | }), |
931 | }), |
932 | rest, |
933 | )) |
934 | } |
935 | |
936 | #[cfg_attr (docsrs, doc(cfg(feature = "parsing" )))] |
937 | impl Parse for LitStr { |
938 | fn parse(input: ParseStream) -> Result<Self> { |
939 | let head = input.fork(); |
940 | match input.parse() { |
941 | Ok(Lit::Str(lit)) => Ok(lit), |
942 | _ => Err(head.error("expected string literal" )), |
943 | } |
944 | } |
945 | } |
946 | |
947 | #[cfg_attr (docsrs, doc(cfg(feature = "parsing" )))] |
948 | impl Parse for LitByteStr { |
949 | fn parse(input: ParseStream) -> Result<Self> { |
950 | let head = input.fork(); |
951 | match input.parse() { |
952 | Ok(Lit::ByteStr(lit)) => Ok(lit), |
953 | _ => Err(head.error("expected byte string literal" )), |
954 | } |
955 | } |
956 | } |
957 | |
958 | #[cfg_attr (docsrs, doc(cfg(feature = "parsing" )))] |
959 | impl Parse for LitCStr { |
960 | fn parse(input: ParseStream) -> Result<Self> { |
961 | let head = input.fork(); |
962 | match input.parse() { |
963 | Ok(Lit::CStr(lit)) => Ok(lit), |
964 | _ => Err(head.error("expected C string literal" )), |
965 | } |
966 | } |
967 | } |
968 | |
969 | #[cfg_attr (docsrs, doc(cfg(feature = "parsing" )))] |
970 | impl Parse for LitByte { |
971 | fn parse(input: ParseStream) -> Result<Self> { |
972 | let head = input.fork(); |
973 | match input.parse() { |
974 | Ok(Lit::Byte(lit)) => Ok(lit), |
975 | _ => Err(head.error("expected byte literal" )), |
976 | } |
977 | } |
978 | } |
979 | |
980 | #[cfg_attr (docsrs, doc(cfg(feature = "parsing" )))] |
981 | impl Parse for LitChar { |
982 | fn parse(input: ParseStream) -> Result<Self> { |
983 | let head = input.fork(); |
984 | match input.parse() { |
985 | Ok(Lit::Char(lit)) => Ok(lit), |
986 | _ => Err(head.error("expected character literal" )), |
987 | } |
988 | } |
989 | } |
990 | |
991 | #[cfg_attr (docsrs, doc(cfg(feature = "parsing" )))] |
992 | impl Parse for LitInt { |
993 | fn parse(input: ParseStream) -> Result<Self> { |
994 | let head = input.fork(); |
995 | match input.parse() { |
996 | Ok(Lit::Int(lit)) => Ok(lit), |
997 | _ => Err(head.error("expected integer literal" )), |
998 | } |
999 | } |
1000 | } |
1001 | |
1002 | #[cfg_attr (docsrs, doc(cfg(feature = "parsing" )))] |
1003 | impl Parse for LitFloat { |
1004 | fn parse(input: ParseStream) -> Result<Self> { |
1005 | let head = input.fork(); |
1006 | match input.parse() { |
1007 | Ok(Lit::Float(lit)) => Ok(lit), |
1008 | _ => Err(head.error("expected floating point literal" )), |
1009 | } |
1010 | } |
1011 | } |
1012 | |
1013 | #[cfg_attr (docsrs, doc(cfg(feature = "parsing" )))] |
1014 | impl Parse for LitBool { |
1015 | fn parse(input: ParseStream) -> Result<Self> { |
1016 | let head = input.fork(); |
1017 | match input.parse() { |
1018 | Ok(Lit::Bool(lit)) => Ok(lit), |
1019 | _ => Err(head.error("expected boolean literal" )), |
1020 | } |
1021 | } |
1022 | } |
1023 | |
1024 | fn peek_impl(cursor: Cursor, peek: fn(ParseStream) -> bool) -> bool { |
1025 | let scope = Span::call_site(); |
1026 | let unexpected = Rc::new(Cell::new(Unexpected::None)); |
1027 | let buffer = crate::parse::new_parse_buffer(scope, cursor, unexpected); |
1028 | peek(&buffer) |
1029 | } |
1030 | |
1031 | macro_rules! impl_token { |
1032 | ($display:literal $name:ty) => { |
1033 | impl Token for $name { |
1034 | fn peek(cursor: Cursor) -> bool { |
1035 | fn peek(input: ParseStream) -> bool { |
1036 | <$name as Parse>::parse(input).is_ok() |
1037 | } |
1038 | peek_impl(cursor, peek) |
1039 | } |
1040 | |
1041 | fn display() -> &'static str { |
1042 | $display |
1043 | } |
1044 | } |
1045 | |
1046 | impl token::private::Sealed for $name {} |
1047 | }; |
1048 | } |
1049 | |
1050 | impl_token!("literal" Lit); |
1051 | impl_token!("string literal" LitStr); |
1052 | impl_token!("byte string literal" LitByteStr); |
1053 | impl_token!("C-string literal" LitCStr); |
1054 | impl_token!("byte literal" LitByte); |
1055 | impl_token!("character literal" LitChar); |
1056 | impl_token!("integer literal" LitInt); |
1057 | impl_token!("floating point literal" LitFloat); |
1058 | impl_token!("boolean literal" LitBool); |
1059 | } |
1060 | |
1061 | #[cfg (feature = "printing" )] |
1062 | mod printing { |
1063 | use crate::lit::{LitBool, LitByte, LitByteStr, LitCStr, LitChar, LitFloat, LitInt, LitStr}; |
1064 | use proc_macro2::TokenStream; |
1065 | use quote::{ToTokens, TokenStreamExt}; |
1066 | |
1067 | #[cfg_attr (docsrs, doc(cfg(feature = "printing" )))] |
1068 | impl ToTokens for LitStr { |
1069 | fn to_tokens(&self, tokens: &mut TokenStream) { |
1070 | self.repr.token.to_tokens(tokens); |
1071 | } |
1072 | } |
1073 | |
1074 | #[cfg_attr (docsrs, doc(cfg(feature = "printing" )))] |
1075 | impl ToTokens for LitByteStr { |
1076 | fn to_tokens(&self, tokens: &mut TokenStream) { |
1077 | self.repr.token.to_tokens(tokens); |
1078 | } |
1079 | } |
1080 | |
1081 | #[cfg_attr (docsrs, doc(cfg(feature = "printing" )))] |
1082 | impl ToTokens for LitCStr { |
1083 | fn to_tokens(&self, tokens: &mut TokenStream) { |
1084 | self.repr.token.to_tokens(tokens); |
1085 | } |
1086 | } |
1087 | |
1088 | #[cfg_attr (docsrs, doc(cfg(feature = "printing" )))] |
1089 | impl ToTokens for LitByte { |
1090 | fn to_tokens(&self, tokens: &mut TokenStream) { |
1091 | self.repr.token.to_tokens(tokens); |
1092 | } |
1093 | } |
1094 | |
1095 | #[cfg_attr (docsrs, doc(cfg(feature = "printing" )))] |
1096 | impl ToTokens for LitChar { |
1097 | fn to_tokens(&self, tokens: &mut TokenStream) { |
1098 | self.repr.token.to_tokens(tokens); |
1099 | } |
1100 | } |
1101 | |
1102 | #[cfg_attr (docsrs, doc(cfg(feature = "printing" )))] |
1103 | impl ToTokens for LitInt { |
1104 | fn to_tokens(&self, tokens: &mut TokenStream) { |
1105 | self.repr.token.to_tokens(tokens); |
1106 | } |
1107 | } |
1108 | |
1109 | #[cfg_attr (docsrs, doc(cfg(feature = "printing" )))] |
1110 | impl ToTokens for LitFloat { |
1111 | fn to_tokens(&self, tokens: &mut TokenStream) { |
1112 | self.repr.token.to_tokens(tokens); |
1113 | } |
1114 | } |
1115 | |
1116 | #[cfg_attr (docsrs, doc(cfg(feature = "printing" )))] |
1117 | impl ToTokens for LitBool { |
1118 | fn to_tokens(&self, tokens: &mut TokenStream) { |
1119 | tokens.append(self.token()); |
1120 | } |
1121 | } |
1122 | } |
1123 | |
1124 | mod value { |
1125 | use crate::bigint::BigInt; |
1126 | use crate::lit::{ |
1127 | Lit, LitBool, LitByte, LitByteStr, LitCStr, LitChar, LitFloat, LitFloatRepr, LitInt, |
1128 | LitIntRepr, LitRepr, LitStr, |
1129 | }; |
1130 | use proc_macro2::{Literal, Span}; |
1131 | use std::ascii; |
1132 | use std::char; |
1133 | use std::ffi::CString; |
1134 | use std::ops::{Index, RangeFrom}; |
1135 | |
1136 | impl Lit { |
1137 | /// Interpret a Syn literal from a proc-macro2 literal. |
1138 | pub fn new(token: Literal) -> Self { |
1139 | let repr = token.to_string(); |
1140 | |
1141 | match byte(&repr, 0) { |
1142 | // "...", r"...", r#"..."# |
1143 | b'"' | b'r' => { |
1144 | let (_, suffix) = parse_lit_str(&repr); |
1145 | return Lit::Str(LitStr { |
1146 | repr: Box::new(LitRepr { token, suffix }), |
1147 | }); |
1148 | } |
1149 | b'b' => match byte(&repr, 1) { |
1150 | // b"...", br"...", br#"...#" |
1151 | b'"' | b'r' => { |
1152 | let (_, suffix) = parse_lit_byte_str(&repr); |
1153 | return Lit::ByteStr(LitByteStr { |
1154 | repr: Box::new(LitRepr { token, suffix }), |
1155 | }); |
1156 | } |
1157 | // b'...' |
1158 | b' \'' => { |
1159 | let (_, suffix) = parse_lit_byte(&repr); |
1160 | return Lit::Byte(LitByte { |
1161 | repr: Box::new(LitRepr { token, suffix }), |
1162 | }); |
1163 | } |
1164 | _ => {} |
1165 | }, |
1166 | // c"...", cr"...", cr#"..."# |
1167 | b'c' => { |
1168 | let (_, suffix) = parse_lit_c_str(&repr); |
1169 | return Lit::CStr(LitCStr { |
1170 | repr: Box::new(LitRepr { token, suffix }), |
1171 | }); |
1172 | } |
1173 | // '...' |
1174 | b' \'' => { |
1175 | let (_, suffix) = parse_lit_char(&repr); |
1176 | return Lit::Char(LitChar { |
1177 | repr: Box::new(LitRepr { token, suffix }), |
1178 | }); |
1179 | } |
1180 | b'0' ..=b'9' | b'-' => { |
1181 | // 0, 123, 0xFF, 0o77, 0b11 |
1182 | if let Some((digits, suffix)) = parse_lit_int(&repr) { |
1183 | return Lit::Int(LitInt { |
1184 | repr: Box::new(LitIntRepr { |
1185 | token, |
1186 | digits, |
1187 | suffix, |
1188 | }), |
1189 | }); |
1190 | } |
1191 | // 1.0, 1e-1, 1e+1 |
1192 | if let Some((digits, suffix)) = parse_lit_float(&repr) { |
1193 | return Lit::Float(LitFloat { |
1194 | repr: Box::new(LitFloatRepr { |
1195 | token, |
1196 | digits, |
1197 | suffix, |
1198 | }), |
1199 | }); |
1200 | } |
1201 | } |
1202 | // true, false |
1203 | b't' | b'f' => { |
1204 | if repr == "true" || repr == "false" { |
1205 | return Lit::Bool(LitBool { |
1206 | value: repr == "true" , |
1207 | span: token.span(), |
1208 | }); |
1209 | } |
1210 | } |
1211 | b'(' if repr == "(/*ERROR*/)" => return Lit::Verbatim(token), |
1212 | _ => {} |
1213 | } |
1214 | |
1215 | panic!("unrecognized literal: ` {}`" , repr); |
1216 | } |
1217 | |
1218 | pub fn suffix(&self) -> &str { |
1219 | match self { |
1220 | Lit::Str(lit) => lit.suffix(), |
1221 | Lit::ByteStr(lit) => lit.suffix(), |
1222 | Lit::CStr(lit) => lit.suffix(), |
1223 | Lit::Byte(lit) => lit.suffix(), |
1224 | Lit::Char(lit) => lit.suffix(), |
1225 | Lit::Int(lit) => lit.suffix(), |
1226 | Lit::Float(lit) => lit.suffix(), |
1227 | Lit::Bool(_) | Lit::Verbatim(_) => "" , |
1228 | } |
1229 | } |
1230 | |
1231 | pub fn span(&self) -> Span { |
1232 | match self { |
1233 | Lit::Str(lit) => lit.span(), |
1234 | Lit::ByteStr(lit) => lit.span(), |
1235 | Lit::CStr(lit) => lit.span(), |
1236 | Lit::Byte(lit) => lit.span(), |
1237 | Lit::Char(lit) => lit.span(), |
1238 | Lit::Int(lit) => lit.span(), |
1239 | Lit::Float(lit) => lit.span(), |
1240 | Lit::Bool(lit) => lit.span, |
1241 | Lit::Verbatim(lit) => lit.span(), |
1242 | } |
1243 | } |
1244 | |
1245 | pub fn set_span(&mut self, span: Span) { |
1246 | match self { |
1247 | Lit::Str(lit) => lit.set_span(span), |
1248 | Lit::ByteStr(lit) => lit.set_span(span), |
1249 | Lit::CStr(lit) => lit.set_span(span), |
1250 | Lit::Byte(lit) => lit.set_span(span), |
1251 | Lit::Char(lit) => lit.set_span(span), |
1252 | Lit::Int(lit) => lit.set_span(span), |
1253 | Lit::Float(lit) => lit.set_span(span), |
1254 | Lit::Bool(lit) => lit.span = span, |
1255 | Lit::Verbatim(lit) => lit.set_span(span), |
1256 | } |
1257 | } |
1258 | } |
1259 | |
1260 | /// Get the byte at offset idx, or a default of `b'\0'` if we're looking |
1261 | /// past the end of the input buffer. |
1262 | pub(crate) fn byte<S: AsRef<[u8]> + ?Sized>(s: &S, idx: usize) -> u8 { |
1263 | let s = s.as_ref(); |
1264 | if idx < s.len() { |
1265 | s[idx] |
1266 | } else { |
1267 | 0 |
1268 | } |
1269 | } |
1270 | |
1271 | fn next_chr(s: &str) -> char { |
1272 | s.chars().next().unwrap_or(' \0' ) |
1273 | } |
1274 | |
1275 | // Returns (content, suffix). |
1276 | pub(crate) fn parse_lit_str(s: &str) -> (Box<str>, Box<str>) { |
1277 | match byte(s, 0) { |
1278 | b'"' => parse_lit_str_cooked(s), |
1279 | b'r' => parse_lit_str_raw(s), |
1280 | _ => unreachable!(), |
1281 | } |
1282 | } |
1283 | |
1284 | fn parse_lit_str_cooked(mut s: &str) -> (Box<str>, Box<str>) { |
1285 | assert_eq!(byte(s, 0), b'"' ); |
1286 | s = &s[1..]; |
1287 | |
1288 | let mut content = String::new(); |
1289 | 'outer: loop { |
1290 | let ch = match byte(s, 0) { |
1291 | b'"' => break, |
1292 | b' \\' => { |
1293 | let b = byte(s, 1); |
1294 | s = &s[2..]; |
1295 | match b { |
1296 | b'x' => { |
1297 | let (byte, rest) = backslash_x(s); |
1298 | s = rest; |
1299 | assert!(byte <= 0x7F, "invalid \\x byte in string literal" ); |
1300 | char::from_u32(u32::from(byte)).unwrap() |
1301 | } |
1302 | b'u' => { |
1303 | let (ch, rest) = backslash_u(s); |
1304 | s = rest; |
1305 | ch |
1306 | } |
1307 | b'n' => ' \n' , |
1308 | b'r' => ' \r' , |
1309 | b't' => ' \t' , |
1310 | b' \\' => ' \\' , |
1311 | b'0' => ' \0' , |
1312 | b' \'' => ' \'' , |
1313 | b'"' => '"' , |
1314 | b' \r' | b' \n' => loop { |
1315 | let b = byte(s, 0); |
1316 | match b { |
1317 | b' ' | b' \t' | b' \n' | b' \r' => s = &s[1..], |
1318 | _ => continue 'outer, |
1319 | } |
1320 | }, |
1321 | b => panic!( |
1322 | "unexpected byte ' {}' after \\ character in string literal" , |
1323 | ascii::escape_default(b), |
1324 | ), |
1325 | } |
1326 | } |
1327 | b' \r' => { |
1328 | assert_eq!(byte(s, 1), b' \n' , "bare CR not allowed in string" ); |
1329 | s = &s[2..]; |
1330 | ' \n' |
1331 | } |
1332 | _ => { |
1333 | let ch = next_chr(s); |
1334 | s = &s[ch.len_utf8()..]; |
1335 | ch |
1336 | } |
1337 | }; |
1338 | content.push(ch); |
1339 | } |
1340 | |
1341 | assert!(s.starts_with('"' )); |
1342 | let content = content.into_boxed_str(); |
1343 | let suffix = s[1..].to_owned().into_boxed_str(); |
1344 | (content, suffix) |
1345 | } |
1346 | |
1347 | fn parse_lit_str_raw(mut s: &str) -> (Box<str>, Box<str>) { |
1348 | assert_eq!(byte(s, 0), b'r' ); |
1349 | s = &s[1..]; |
1350 | |
1351 | let mut pounds = 0; |
1352 | while byte(s, pounds) == b'#' { |
1353 | pounds += 1; |
1354 | } |
1355 | assert_eq!(byte(s, pounds), b'"' ); |
1356 | let close = s.rfind('"' ).unwrap(); |
1357 | for end in s[close + 1..close + 1 + pounds].bytes() { |
1358 | assert_eq!(end, b'#' ); |
1359 | } |
1360 | |
1361 | let content = s[pounds + 1..close].to_owned().into_boxed_str(); |
1362 | let suffix = s[close + 1 + pounds..].to_owned().into_boxed_str(); |
1363 | (content, suffix) |
1364 | } |
1365 | |
1366 | // Returns (content, suffix). |
1367 | pub(crate) fn parse_lit_byte_str(s: &str) -> (Vec<u8>, Box<str>) { |
1368 | assert_eq!(byte(s, 0), b'b' ); |
1369 | match byte(s, 1) { |
1370 | b'"' => parse_lit_byte_str_cooked(s), |
1371 | b'r' => parse_lit_byte_str_raw(s), |
1372 | _ => unreachable!(), |
1373 | } |
1374 | } |
1375 | |
1376 | fn parse_lit_byte_str_cooked(mut s: &str) -> (Vec<u8>, Box<str>) { |
1377 | assert_eq!(byte(s, 0), b'b' ); |
1378 | assert_eq!(byte(s, 1), b'"' ); |
1379 | s = &s[2..]; |
1380 | |
1381 | // We're going to want to have slices which don't respect codepoint boundaries. |
1382 | let mut v = s.as_bytes(); |
1383 | |
1384 | let mut out = Vec::new(); |
1385 | 'outer: loop { |
1386 | let byte = match byte(v, 0) { |
1387 | b'"' => break, |
1388 | b' \\' => { |
1389 | let b = byte(v, 1); |
1390 | v = &v[2..]; |
1391 | match b { |
1392 | b'x' => { |
1393 | let (b, rest) = backslash_x(v); |
1394 | v = rest; |
1395 | b |
1396 | } |
1397 | b'n' => b' \n' , |
1398 | b'r' => b' \r' , |
1399 | b't' => b' \t' , |
1400 | b' \\' => b' \\' , |
1401 | b'0' => b' \0' , |
1402 | b' \'' => b' \'' , |
1403 | b'"' => b'"' , |
1404 | b' \r' | b' \n' => loop { |
1405 | let byte = byte(v, 0); |
1406 | if matches!(byte, b' ' | b' \t' | b' \n' | b' \r' ) { |
1407 | v = &v[1..]; |
1408 | } else { |
1409 | continue 'outer; |
1410 | } |
1411 | }, |
1412 | b => panic!( |
1413 | "unexpected byte ' {}' after \\ character in byte-string literal" , |
1414 | ascii::escape_default(b), |
1415 | ), |
1416 | } |
1417 | } |
1418 | b' \r' => { |
1419 | assert_eq!(byte(v, 1), b' \n' , "bare CR not allowed in string" ); |
1420 | v = &v[2..]; |
1421 | b' \n' |
1422 | } |
1423 | b => { |
1424 | v = &v[1..]; |
1425 | b |
1426 | } |
1427 | }; |
1428 | out.push(byte); |
1429 | } |
1430 | |
1431 | assert_eq!(byte(v, 0), b'"' ); |
1432 | let suffix = s[s.len() - v.len() + 1..].to_owned().into_boxed_str(); |
1433 | (out, suffix) |
1434 | } |
1435 | |
1436 | fn parse_lit_byte_str_raw(s: &str) -> (Vec<u8>, Box<str>) { |
1437 | assert_eq!(byte(s, 0), b'b' ); |
1438 | let (value, suffix) = parse_lit_str_raw(&s[1..]); |
1439 | (String::from(value).into_bytes(), suffix) |
1440 | } |
1441 | |
1442 | // Returns (content, suffix). |
1443 | pub(crate) fn parse_lit_c_str(s: &str) -> (CString, Box<str>) { |
1444 | assert_eq!(byte(s, 0), b'c' ); |
1445 | match byte(s, 1) { |
1446 | b'"' => parse_lit_c_str_cooked(s), |
1447 | b'r' => parse_lit_c_str_raw(s), |
1448 | _ => unreachable!(), |
1449 | } |
1450 | } |
1451 | |
1452 | fn parse_lit_c_str_cooked(mut s: &str) -> (CString, Box<str>) { |
1453 | assert_eq!(byte(s, 0), b'c' ); |
1454 | assert_eq!(byte(s, 1), b'"' ); |
1455 | s = &s[2..]; |
1456 | |
1457 | // We're going to want to have slices which don't respect codepoint boundaries. |
1458 | let mut v = s.as_bytes(); |
1459 | |
1460 | let mut out = Vec::new(); |
1461 | 'outer: loop { |
1462 | let byte = match byte(v, 0) { |
1463 | b'"' => break, |
1464 | b' \\' => { |
1465 | let b = byte(v, 1); |
1466 | v = &v[2..]; |
1467 | match b { |
1468 | b'x' => { |
1469 | let (b, rest) = backslash_x(v); |
1470 | assert!(b != 0, " \\x00 is not allowed in C-string literal" ); |
1471 | v = rest; |
1472 | b |
1473 | } |
1474 | b'u' => { |
1475 | let (ch, rest) = backslash_u(v); |
1476 | assert!(ch != ' \0' , " \\u {{0 }} is not allowed in C-string literal" ); |
1477 | v = rest; |
1478 | out.extend_from_slice(ch.encode_utf8(&mut [0u8; 4]).as_bytes()); |
1479 | continue 'outer; |
1480 | } |
1481 | b'n' => b' \n' , |
1482 | b'r' => b' \r' , |
1483 | b't' => b' \t' , |
1484 | b' \\' => b' \\' , |
1485 | b' \'' => b' \'' , |
1486 | b'"' => b'"' , |
1487 | b' \r' | b' \n' => loop { |
1488 | let byte = byte(v, 0); |
1489 | if matches!(byte, b' ' | b' \t' | b' \n' | b' \r' ) { |
1490 | v = &v[1..]; |
1491 | } else { |
1492 | continue 'outer; |
1493 | } |
1494 | }, |
1495 | b => panic!( |
1496 | "unexpected byte ' {}' after \\ character in byte literal" , |
1497 | ascii::escape_default(b), |
1498 | ), |
1499 | } |
1500 | } |
1501 | b' \r' => { |
1502 | assert_eq!(byte(v, 1), b' \n' , "bare CR not allowed in string" ); |
1503 | v = &v[2..]; |
1504 | b' \n' |
1505 | } |
1506 | b => { |
1507 | v = &v[1..]; |
1508 | b |
1509 | } |
1510 | }; |
1511 | out.push(byte); |
1512 | } |
1513 | |
1514 | assert_eq!(byte(v, 0), b'"' ); |
1515 | let suffix = s[s.len() - v.len() + 1..].to_owned().into_boxed_str(); |
1516 | (CString::new(out).unwrap(), suffix) |
1517 | } |
1518 | |
1519 | fn parse_lit_c_str_raw(s: &str) -> (CString, Box<str>) { |
1520 | assert_eq!(byte(s, 0), b'c' ); |
1521 | let (value, suffix) = parse_lit_str_raw(&s[1..]); |
1522 | (CString::new(String::from(value)).unwrap(), suffix) |
1523 | } |
1524 | |
1525 | // Returns (value, suffix). |
1526 | pub(crate) fn parse_lit_byte(s: &str) -> (u8, Box<str>) { |
1527 | assert_eq!(byte(s, 0), b'b' ); |
1528 | assert_eq!(byte(s, 1), b' \'' ); |
1529 | |
1530 | // We're going to want to have slices which don't respect codepoint boundaries. |
1531 | let mut v = &s.as_bytes()[2..]; |
1532 | |
1533 | let b = match byte(v, 0) { |
1534 | b' \\' => { |
1535 | let b = byte(v, 1); |
1536 | v = &v[2..]; |
1537 | match b { |
1538 | b'x' => { |
1539 | let (b, rest) = backslash_x(v); |
1540 | v = rest; |
1541 | b |
1542 | } |
1543 | b'n' => b' \n' , |
1544 | b'r' => b' \r' , |
1545 | b't' => b' \t' , |
1546 | b' \\' => b' \\' , |
1547 | b'0' => b' \0' , |
1548 | b' \'' => b' \'' , |
1549 | b'"' => b'"' , |
1550 | b => panic!( |
1551 | "unexpected byte ' {}' after \\ character in byte literal" , |
1552 | ascii::escape_default(b), |
1553 | ), |
1554 | } |
1555 | } |
1556 | b => { |
1557 | v = &v[1..]; |
1558 | b |
1559 | } |
1560 | }; |
1561 | |
1562 | assert_eq!(byte(v, 0), b' \'' ); |
1563 | let suffix = s[s.len() - v.len() + 1..].to_owned().into_boxed_str(); |
1564 | (b, suffix) |
1565 | } |
1566 | |
1567 | // Returns (value, suffix). |
1568 | pub(crate) fn parse_lit_char(mut s: &str) -> (char, Box<str>) { |
1569 | assert_eq!(byte(s, 0), b' \'' ); |
1570 | s = &s[1..]; |
1571 | |
1572 | let ch = match byte(s, 0) { |
1573 | b' \\' => { |
1574 | let b = byte(s, 1); |
1575 | s = &s[2..]; |
1576 | match b { |
1577 | b'x' => { |
1578 | let (byte, rest) = backslash_x(s); |
1579 | s = rest; |
1580 | assert!(byte <= 0x7F, "invalid \\x byte in character literal" ); |
1581 | char::from_u32(u32::from(byte)).unwrap() |
1582 | } |
1583 | b'u' => { |
1584 | let (ch, rest) = backslash_u(s); |
1585 | s = rest; |
1586 | ch |
1587 | } |
1588 | b'n' => ' \n' , |
1589 | b'r' => ' \r' , |
1590 | b't' => ' \t' , |
1591 | b' \\' => ' \\' , |
1592 | b'0' => ' \0' , |
1593 | b' \'' => ' \'' , |
1594 | b'"' => '"' , |
1595 | b => panic!( |
1596 | "unexpected byte ' {}' after \\ character in character literal" , |
1597 | ascii::escape_default(b), |
1598 | ), |
1599 | } |
1600 | } |
1601 | _ => { |
1602 | let ch = next_chr(s); |
1603 | s = &s[ch.len_utf8()..]; |
1604 | ch |
1605 | } |
1606 | }; |
1607 | assert_eq!(byte(s, 0), b' \'' ); |
1608 | let suffix = s[1..].to_owned().into_boxed_str(); |
1609 | (ch, suffix) |
1610 | } |
1611 | |
1612 | fn backslash_x<S>(s: &S) -> (u8, &S) |
1613 | where |
1614 | S: Index<RangeFrom<usize>, Output = S> + AsRef<[u8]> + ?Sized, |
1615 | { |
1616 | let mut ch = 0; |
1617 | let b0 = byte(s, 0); |
1618 | let b1 = byte(s, 1); |
1619 | ch += 0x10 |
1620 | * match b0 { |
1621 | b'0' ..=b'9' => b0 - b'0' , |
1622 | b'a' ..=b'f' => 10 + (b0 - b'a' ), |
1623 | b'A' ..=b'F' => 10 + (b0 - b'A' ), |
1624 | _ => panic!("unexpected non-hex character after \\x" ), |
1625 | }; |
1626 | ch += match b1 { |
1627 | b'0' ..=b'9' => b1 - b'0' , |
1628 | b'a' ..=b'f' => 10 + (b1 - b'a' ), |
1629 | b'A' ..=b'F' => 10 + (b1 - b'A' ), |
1630 | _ => panic!("unexpected non-hex character after \\x" ), |
1631 | }; |
1632 | (ch, &s[2..]) |
1633 | } |
1634 | |
1635 | fn backslash_u<S>(mut s: &S) -> (char, &S) |
1636 | where |
1637 | S: Index<RangeFrom<usize>, Output = S> + AsRef<[u8]> + ?Sized, |
1638 | { |
1639 | if byte(s, 0) != b'{' { |
1640 | panic!("{}" , "expected { after \\u" ); |
1641 | } |
1642 | s = &s[1..]; |
1643 | |
1644 | let mut ch = 0; |
1645 | let mut digits = 0; |
1646 | loop { |
1647 | let b = byte(s, 0); |
1648 | let digit = match b { |
1649 | b'0' ..=b'9' => b - b'0' , |
1650 | b'a' ..=b'f' => 10 + b - b'a' , |
1651 | b'A' ..=b'F' => 10 + b - b'A' , |
1652 | b'_' if digits > 0 => { |
1653 | s = &s[1..]; |
1654 | continue; |
1655 | } |
1656 | b'}' if digits == 0 => panic!("invalid empty unicode escape" ), |
1657 | b'}' => break, |
1658 | _ => panic!("unexpected non-hex character after \\u" ), |
1659 | }; |
1660 | if digits == 6 { |
1661 | panic!("overlong unicode escape (must have at most 6 hex digits)" ); |
1662 | } |
1663 | ch *= 0x10; |
1664 | ch += u32::from(digit); |
1665 | digits += 1; |
1666 | s = &s[1..]; |
1667 | } |
1668 | assert!(byte(s, 0) == b'}' ); |
1669 | s = &s[1..]; |
1670 | |
1671 | if let Some(ch) = char::from_u32(ch) { |
1672 | (ch, s) |
1673 | } else { |
1674 | panic!("character code {:x} is not a valid unicode character" , ch); |
1675 | } |
1676 | } |
1677 | |
1678 | // Returns base 10 digits and suffix. |
1679 | pub(crate) fn parse_lit_int(mut s: &str) -> Option<(Box<str>, Box<str>)> { |
1680 | let negative = byte(s, 0) == b'-' ; |
1681 | if negative { |
1682 | s = &s[1..]; |
1683 | } |
1684 | |
1685 | let base = match (byte(s, 0), byte(s, 1)) { |
1686 | (b'0' , b'x' ) => { |
1687 | s = &s[2..]; |
1688 | 16 |
1689 | } |
1690 | (b'0' , b'o' ) => { |
1691 | s = &s[2..]; |
1692 | 8 |
1693 | } |
1694 | (b'0' , b'b' ) => { |
1695 | s = &s[2..]; |
1696 | 2 |
1697 | } |
1698 | (b'0' ..=b'9' , _) => 10, |
1699 | _ => return None, |
1700 | }; |
1701 | |
1702 | let mut value = BigInt::new(); |
1703 | let mut has_digit = false; |
1704 | 'outer: loop { |
1705 | let b = byte(s, 0); |
1706 | let digit = match b { |
1707 | b'0' ..=b'9' => b - b'0' , |
1708 | b'a' ..=b'f' if base > 10 => b - b'a' + 10, |
1709 | b'A' ..=b'F' if base > 10 => b - b'A' + 10, |
1710 | b'_' => { |
1711 | s = &s[1..]; |
1712 | continue; |
1713 | } |
1714 | // If looking at a floating point literal, we don't want to |
1715 | // consider it an integer. |
1716 | b'.' if base == 10 => return None, |
1717 | b'e' | b'E' if base == 10 => { |
1718 | let mut has_exp = false; |
1719 | for (i, b) in s[1..].bytes().enumerate() { |
1720 | match b { |
1721 | b'_' => {} |
1722 | b'-' | b'+' => return None, |
1723 | b'0' ..=b'9' => has_exp = true, |
1724 | _ => { |
1725 | let suffix = &s[1 + i..]; |
1726 | if has_exp && crate::ident::xid_ok(suffix) { |
1727 | return None; |
1728 | } else { |
1729 | break 'outer; |
1730 | } |
1731 | } |
1732 | } |
1733 | } |
1734 | if has_exp { |
1735 | return None; |
1736 | } else { |
1737 | break; |
1738 | } |
1739 | } |
1740 | _ => break, |
1741 | }; |
1742 | |
1743 | if digit >= base { |
1744 | return None; |
1745 | } |
1746 | |
1747 | has_digit = true; |
1748 | value *= base; |
1749 | value += digit; |
1750 | s = &s[1..]; |
1751 | } |
1752 | |
1753 | if !has_digit { |
1754 | return None; |
1755 | } |
1756 | |
1757 | let suffix = s; |
1758 | if suffix.is_empty() || crate::ident::xid_ok(suffix) { |
1759 | let mut repr = value.to_string(); |
1760 | if negative { |
1761 | repr.insert(0, '-' ); |
1762 | } |
1763 | Some((repr.into_boxed_str(), suffix.to_owned().into_boxed_str())) |
1764 | } else { |
1765 | None |
1766 | } |
1767 | } |
1768 | |
1769 | // Returns base 10 digits and suffix. |
1770 | pub(crate) fn parse_lit_float(input: &str) -> Option<(Box<str>, Box<str>)> { |
1771 | // Rust's floating point literals are very similar to the ones parsed by |
1772 | // the standard library, except that rust's literals can contain |
1773 | // ignorable underscores. Let's remove those underscores. |
1774 | |
1775 | let mut bytes = input.to_owned().into_bytes(); |
1776 | |
1777 | let start = (*bytes.first()? == b'-' ) as usize; |
1778 | match bytes.get(start)? { |
1779 | b'0' ..=b'9' => {} |
1780 | _ => return None, |
1781 | } |
1782 | |
1783 | let mut read = start; |
1784 | let mut write = start; |
1785 | let mut has_dot = false; |
1786 | let mut has_e = false; |
1787 | let mut has_sign = false; |
1788 | let mut has_exponent = false; |
1789 | while read < bytes.len() { |
1790 | match bytes[read] { |
1791 | b'_' => { |
1792 | // Don't increase write |
1793 | read += 1; |
1794 | continue; |
1795 | } |
1796 | b'0' ..=b'9' => { |
1797 | if has_e { |
1798 | has_exponent = true; |
1799 | } |
1800 | bytes[write] = bytes[read]; |
1801 | } |
1802 | b'.' => { |
1803 | if has_e || has_dot { |
1804 | return None; |
1805 | } |
1806 | has_dot = true; |
1807 | bytes[write] = b'.' ; |
1808 | } |
1809 | b'e' | b'E' => { |
1810 | match bytes[read + 1..] |
1811 | .iter() |
1812 | .find(|b| **b != b'_' ) |
1813 | .unwrap_or(&b' \0' ) |
1814 | { |
1815 | b'-' | b'+' | b'0' ..=b'9' => {} |
1816 | _ => break, |
1817 | } |
1818 | if has_e { |
1819 | if has_exponent { |
1820 | break; |
1821 | } else { |
1822 | return None; |
1823 | } |
1824 | } |
1825 | has_e = true; |
1826 | bytes[write] = b'e' ; |
1827 | } |
1828 | b'-' | b'+' => { |
1829 | if has_sign || has_exponent || !has_e { |
1830 | return None; |
1831 | } |
1832 | has_sign = true; |
1833 | if bytes[read] == b'-' { |
1834 | bytes[write] = bytes[read]; |
1835 | } else { |
1836 | // Omit '+' |
1837 | read += 1; |
1838 | continue; |
1839 | } |
1840 | } |
1841 | _ => break, |
1842 | } |
1843 | read += 1; |
1844 | write += 1; |
1845 | } |
1846 | |
1847 | if has_e && !has_exponent { |
1848 | return None; |
1849 | } |
1850 | |
1851 | let mut digits = String::from_utf8(bytes).unwrap(); |
1852 | let suffix = digits.split_off(read); |
1853 | digits.truncate(write); |
1854 | if suffix.is_empty() || crate::ident::xid_ok(&suffix) { |
1855 | Some((digits.into_boxed_str(), suffix.into_boxed_str())) |
1856 | } else { |
1857 | None |
1858 | } |
1859 | } |
1860 | } |
1861 | |