1 | //! Stream capability for combinators to parse |
2 | //! |
3 | //! Stream types include: |
4 | //! - `&[u8]` and [`Bytes`] for binary data |
5 | //! - `&str` (aliased as [`Str`]) and [`BStr`] for UTF-8 data |
6 | //! - [`LocatingSlice`] can track the location within the original buffer to report |
7 | //! [spans][crate::Parser::with_span] |
8 | //! - [`Stateful`] to thread global state through your parsers |
9 | //! - [`Partial`] can mark an input as partial buffer that is being streamed into |
10 | //! - [Custom stream types][crate::_topic::stream] |
11 | |
12 | use core::hash::BuildHasher; |
13 | use core::num::NonZeroUsize; |
14 | |
15 | use crate::ascii::Caseless as AsciiCaseless; |
16 | use crate::error::Needed; |
17 | use crate::lib::std::iter::{Cloned, Enumerate}; |
18 | use crate::lib::std::slice::Iter; |
19 | use crate::lib::std::str::from_utf8; |
20 | use crate::lib::std::str::CharIndices; |
21 | use crate::lib::std::str::FromStr; |
22 | |
23 | #[allow (unused_imports)] |
24 | #[cfg (any(feature = "unstable-doc" , feature = "unstable-recover" ))] |
25 | use crate::error::ErrMode; |
26 | |
27 | #[cfg (feature = "alloc" )] |
28 | use crate::lib::std::collections::BTreeMap; |
29 | #[cfg (feature = "alloc" )] |
30 | use crate::lib::std::collections::BTreeSet; |
31 | #[cfg (feature = "std" )] |
32 | use crate::lib::std::collections::HashMap; |
33 | #[cfg (feature = "std" )] |
34 | use crate::lib::std::collections::HashSet; |
35 | #[cfg (feature = "alloc" )] |
36 | use crate::lib::std::string::String; |
37 | #[cfg (feature = "alloc" )] |
38 | use crate::lib::std::vec::Vec; |
39 | |
40 | mod bstr; |
41 | mod bytes; |
42 | mod locating; |
43 | mod partial; |
44 | mod range; |
45 | #[cfg (feature = "unstable-recover" )] |
46 | #[cfg (feature = "std" )] |
47 | mod recoverable; |
48 | mod stateful; |
49 | #[cfg (test)] |
50 | mod tests; |
51 | mod token; |
52 | |
53 | pub use bstr::BStr; |
54 | pub use bytes::Bytes; |
55 | pub use locating::LocatingSlice; |
56 | pub use partial::Partial; |
57 | pub use range::Range; |
58 | #[cfg (feature = "unstable-recover" )] |
59 | #[cfg (feature = "std" )] |
60 | pub use recoverable::Recoverable; |
61 | pub use stateful::Stateful; |
62 | pub use token::TokenSlice; |
63 | |
64 | /// UTF-8 Stream |
65 | pub type Str<'i> = &'i str; |
66 | |
67 | /// Abstract method to calculate the input length |
68 | pub trait SliceLen { |
69 | /// Calculates the input length, as indicated by its name, |
70 | /// and the name of the trait itself |
71 | fn slice_len(&self) -> usize; |
72 | } |
73 | |
74 | impl<S: SliceLen> SliceLen for AsciiCaseless<S> { |
75 | #[inline (always)] |
76 | fn slice_len(&self) -> usize { |
77 | self.0.slice_len() |
78 | } |
79 | } |
80 | |
81 | impl<T> SliceLen for &[T] { |
82 | #[inline (always)] |
83 | fn slice_len(&self) -> usize { |
84 | self.len() |
85 | } |
86 | } |
87 | |
88 | impl<T, const LEN: usize> SliceLen for [T; LEN] { |
89 | #[inline (always)] |
90 | fn slice_len(&self) -> usize { |
91 | self.len() |
92 | } |
93 | } |
94 | |
95 | impl<T, const LEN: usize> SliceLen for &[T; LEN] { |
96 | #[inline (always)] |
97 | fn slice_len(&self) -> usize { |
98 | self.len() |
99 | } |
100 | } |
101 | |
102 | impl SliceLen for &str { |
103 | #[inline (always)] |
104 | fn slice_len(&self) -> usize { |
105 | self.len() |
106 | } |
107 | } |
108 | |
109 | impl SliceLen for u8 { |
110 | #[inline (always)] |
111 | fn slice_len(&self) -> usize { |
112 | 1 |
113 | } |
114 | } |
115 | |
116 | impl SliceLen for char { |
117 | #[inline (always)] |
118 | fn slice_len(&self) -> usize { |
119 | self.len_utf8() |
120 | } |
121 | } |
122 | |
123 | impl<I> SliceLen for (I, usize, usize) |
124 | where |
125 | I: SliceLen, |
126 | { |
127 | #[inline (always)] |
128 | fn slice_len(&self) -> usize { |
129 | self.0.slice_len() * 8 + self.2 - self.1 |
130 | } |
131 | } |
132 | |
133 | /// Core definition for parser input state |
134 | pub trait Stream: Offset<<Self as Stream>::Checkpoint> + crate::lib::std::fmt::Debug { |
135 | /// The smallest unit being parsed |
136 | /// |
137 | /// Example: `u8` for `&[u8]` or `char` for `&str` |
138 | type Token: crate::lib::std::fmt::Debug; |
139 | /// Sequence of `Token`s |
140 | /// |
141 | /// Example: `&[u8]` for `LocatingSlice<&[u8]>` or `&str` for `LocatingSlice<&str>` |
142 | type Slice: crate::lib::std::fmt::Debug; |
143 | |
144 | /// Iterate with the offset from the current location |
145 | type IterOffsets: Iterator<Item = (usize, Self::Token)>; |
146 | |
147 | /// A parse location within the stream |
148 | type Checkpoint: Offset + Clone + crate::lib::std::fmt::Debug; |
149 | |
150 | /// Iterate with the offset from the current location |
151 | fn iter_offsets(&self) -> Self::IterOffsets; |
152 | |
153 | /// Returns the offset to the end of the input |
154 | fn eof_offset(&self) -> usize; |
155 | |
156 | /// Split off the next token from the input |
157 | fn next_token(&mut self) -> Option<Self::Token>; |
158 | /// Split off the next token from the input |
159 | fn peek_token(&self) -> Option<Self::Token>; |
160 | |
161 | /// Finds the offset of the next matching token |
162 | fn offset_for<P>(&self, predicate: P) -> Option<usize> |
163 | where |
164 | P: Fn(Self::Token) -> bool; |
165 | /// Get the offset for the number of `tokens` into the stream |
166 | /// |
167 | /// This means "0 tokens" will return `0` offset |
168 | fn offset_at(&self, tokens: usize) -> Result<usize, Needed>; |
169 | /// Split off a slice of tokens from the input |
170 | /// |
171 | /// <div class="warning"> |
172 | /// |
173 | /// **Note:** For inputs with variable width tokens, like `&str`'s `char`, `offset` might not correspond |
174 | /// with the number of tokens. To get a valid offset, use: |
175 | /// - [`Stream::eof_offset`] |
176 | /// - [`Stream::iter_offsets`] |
177 | /// - [`Stream::offset_for`] |
178 | /// - [`Stream::offset_at`] |
179 | /// |
180 | /// </div> |
181 | /// |
182 | /// # Panic |
183 | /// |
184 | /// This will panic if |
185 | /// |
186 | /// * Indexes must be within bounds of the original input; |
187 | /// * Indexes must uphold invariants of the stream, like for `str` they must lie on UTF-8 |
188 | /// sequence boundaries. |
189 | /// |
190 | fn next_slice(&mut self, offset: usize) -> Self::Slice; |
191 | /// Split off a slice of tokens from the input |
192 | /// |
193 | /// <div class="warning"> |
194 | /// |
195 | /// **Note:** For inputs with variable width tokens, like `&str`'s `char`, `offset` might not correspond |
196 | /// with the number of tokens. To get a valid offset, use: |
197 | /// - [`Stream::eof_offset`] |
198 | /// - [`Stream::iter_offsets`] |
199 | /// - [`Stream::offset_for`] |
200 | /// - [`Stream::offset_at`] |
201 | /// |
202 | /// </div> |
203 | /// |
204 | /// # Safety |
205 | /// |
206 | /// Callers of this function are responsible that these preconditions are satisfied: |
207 | /// |
208 | /// * Indexes must be within bounds of the original input; |
209 | /// * Indexes must uphold invariants of the stream, like for `str` they must lie on UTF-8 |
210 | /// sequence boundaries. |
211 | /// |
212 | unsafe fn next_slice_unchecked(&mut self, offset: usize) -> Self::Slice { |
213 | // Inherent impl to allow callers to have `unsafe`-free code |
214 | self.next_slice(offset) |
215 | } |
216 | /// Split off a slice of tokens from the input |
217 | fn peek_slice(&self, offset: usize) -> Self::Slice; |
218 | /// Split off a slice of tokens from the input |
219 | /// |
220 | /// # Safety |
221 | /// |
222 | /// Callers of this function are responsible that these preconditions are satisfied: |
223 | /// |
224 | /// * Indexes must be within bounds of the original input; |
225 | /// * Indexes must uphold invariants of the stream, like for `str` they must lie on UTF-8 |
226 | /// sequence boundaries. |
227 | unsafe fn peek_slice_unchecked(&self, offset: usize) -> Self::Slice { |
228 | // Inherent impl to allow callers to have `unsafe`-free code |
229 | self.peek_slice(offset) |
230 | } |
231 | |
232 | /// Advance to the end of the stream |
233 | #[inline (always)] |
234 | fn finish(&mut self) -> Self::Slice { |
235 | self.next_slice(self.eof_offset()) |
236 | } |
237 | /// Advance to the end of the stream |
238 | #[inline (always)] |
239 | fn peek_finish(&self) -> Self::Slice |
240 | where |
241 | Self: Clone, |
242 | { |
243 | self.peek_slice(self.eof_offset()) |
244 | } |
245 | |
246 | /// Save the current parse location within the stream |
247 | fn checkpoint(&self) -> Self::Checkpoint; |
248 | /// Revert the stream to a prior [`Self::Checkpoint`] |
249 | /// |
250 | /// # Panic |
251 | /// |
252 | /// May panic if an invalid [`Self::Checkpoint`] is provided |
253 | fn reset(&mut self, checkpoint: &Self::Checkpoint); |
254 | |
255 | /// Return the inner-most stream |
256 | fn raw(&self) -> &dyn crate::lib::std::fmt::Debug; |
257 | } |
258 | |
259 | impl<'i, T> Stream for &'i [T] |
260 | where |
261 | T: Clone + crate::lib::std::fmt::Debug, |
262 | { |
263 | type Token = T; |
264 | type Slice = &'i [T]; |
265 | |
266 | type IterOffsets = Enumerate<Cloned<Iter<'i, T>>>; |
267 | |
268 | type Checkpoint = Checkpoint<Self, Self>; |
269 | |
270 | #[inline (always)] |
271 | fn iter_offsets(&self) -> Self::IterOffsets { |
272 | self.iter().cloned().enumerate() |
273 | } |
274 | #[inline (always)] |
275 | fn eof_offset(&self) -> usize { |
276 | self.len() |
277 | } |
278 | |
279 | #[inline (always)] |
280 | fn next_token(&mut self) -> Option<Self::Token> { |
281 | let (token, next) = self.split_first()?; |
282 | *self = next; |
283 | Some(token.clone()) |
284 | } |
285 | |
286 | #[inline (always)] |
287 | fn peek_token(&self) -> Option<Self::Token> { |
288 | if self.is_empty() { |
289 | None |
290 | } else { |
291 | Some(self[0].clone()) |
292 | } |
293 | } |
294 | |
295 | #[inline (always)] |
296 | fn offset_for<P>(&self, predicate: P) -> Option<usize> |
297 | where |
298 | P: Fn(Self::Token) -> bool, |
299 | { |
300 | self.iter().position(|b| predicate(b.clone())) |
301 | } |
302 | #[inline (always)] |
303 | fn offset_at(&self, tokens: usize) -> Result<usize, Needed> { |
304 | if let Some(needed) = tokens.checked_sub(self.len()).and_then(NonZeroUsize::new) { |
305 | Err(Needed::Size(needed)) |
306 | } else { |
307 | Ok(tokens) |
308 | } |
309 | } |
310 | #[inline (always)] |
311 | fn next_slice(&mut self, offset: usize) -> Self::Slice { |
312 | let (slice, next) = self.split_at(offset); |
313 | *self = next; |
314 | slice |
315 | } |
316 | #[inline (always)] |
317 | unsafe fn next_slice_unchecked(&mut self, offset: usize) -> Self::Slice { |
318 | #[cfg (debug_assertions)] |
319 | self.peek_slice(offset); |
320 | |
321 | // SAFETY: `Stream::next_slice_unchecked` requires `offset` to be in bounds |
322 | let slice = unsafe { self.get_unchecked(..offset) }; |
323 | // SAFETY: `Stream::next_slice_unchecked` requires `offset` to be in bounds |
324 | let next = unsafe { self.get_unchecked(offset..) }; |
325 | *self = next; |
326 | slice |
327 | } |
328 | #[inline (always)] |
329 | fn peek_slice(&self, offset: usize) -> Self::Slice { |
330 | &self[..offset] |
331 | } |
332 | #[inline (always)] |
333 | unsafe fn peek_slice_unchecked(&self, offset: usize) -> Self::Slice { |
334 | #[cfg (debug_assertions)] |
335 | self.peek_slice(offset); |
336 | |
337 | // SAFETY: `Stream::next_slice_unchecked` requires `offset` to be in bounds |
338 | let slice = unsafe { self.get_unchecked(..offset) }; |
339 | slice |
340 | } |
341 | |
342 | #[inline (always)] |
343 | fn checkpoint(&self) -> Self::Checkpoint { |
344 | Checkpoint::<_, Self>::new(*self) |
345 | } |
346 | #[inline (always)] |
347 | fn reset(&mut self, checkpoint: &Self::Checkpoint) { |
348 | *self = checkpoint.inner; |
349 | } |
350 | |
351 | #[inline (always)] |
352 | fn raw(&self) -> &dyn crate::lib::std::fmt::Debug { |
353 | self |
354 | } |
355 | } |
356 | |
357 | impl<'i> Stream for &'i str { |
358 | type Token = char; |
359 | type Slice = &'i str; |
360 | |
361 | type IterOffsets = CharIndices<'i>; |
362 | |
363 | type Checkpoint = Checkpoint<Self, Self>; |
364 | |
365 | #[inline (always)] |
366 | fn iter_offsets(&self) -> Self::IterOffsets { |
367 | self.char_indices() |
368 | } |
369 | #[inline (always)] |
370 | fn eof_offset(&self) -> usize { |
371 | self.len() |
372 | } |
373 | |
374 | #[inline (always)] |
375 | fn next_token(&mut self) -> Option<Self::Token> { |
376 | let c = self.chars().next()?; |
377 | let offset = c.len(); |
378 | *self = &self[offset..]; |
379 | Some(c) |
380 | } |
381 | |
382 | #[inline (always)] |
383 | fn peek_token(&self) -> Option<Self::Token> { |
384 | self.chars().next() |
385 | } |
386 | |
387 | #[inline (always)] |
388 | fn offset_for<P>(&self, predicate: P) -> Option<usize> |
389 | where |
390 | P: Fn(Self::Token) -> bool, |
391 | { |
392 | for (o, c) in self.iter_offsets() { |
393 | if predicate(c) { |
394 | return Some(o); |
395 | } |
396 | } |
397 | None |
398 | } |
399 | #[inline ] |
400 | fn offset_at(&self, tokens: usize) -> Result<usize, Needed> { |
401 | let mut cnt = 0; |
402 | for (offset, _) in self.iter_offsets() { |
403 | if cnt == tokens { |
404 | return Ok(offset); |
405 | } |
406 | cnt += 1; |
407 | } |
408 | |
409 | if cnt == tokens { |
410 | Ok(self.eof_offset()) |
411 | } else { |
412 | Err(Needed::Unknown) |
413 | } |
414 | } |
415 | #[inline (always)] |
416 | fn next_slice(&mut self, offset: usize) -> Self::Slice { |
417 | let (slice, next) = self.split_at(offset); |
418 | *self = next; |
419 | slice |
420 | } |
421 | #[inline (always)] |
422 | unsafe fn next_slice_unchecked(&mut self, offset: usize) -> Self::Slice { |
423 | #[cfg (debug_assertions)] |
424 | self.peek_slice(offset); |
425 | |
426 | // SAFETY: `Stream::next_slice_unchecked` requires `offset` to be in bounds and on a UTF-8 |
427 | // sequence boundary |
428 | let slice = unsafe { self.get_unchecked(..offset) }; |
429 | // SAFETY: `Stream::next_slice_unchecked` requires `offset` to be in bounds and on a UTF-8 |
430 | // sequence boundary |
431 | let next = unsafe { self.get_unchecked(offset..) }; |
432 | *self = next; |
433 | slice |
434 | } |
435 | #[inline (always)] |
436 | fn peek_slice(&self, offset: usize) -> Self::Slice { |
437 | &self[..offset] |
438 | } |
439 | #[inline (always)] |
440 | unsafe fn peek_slice_unchecked(&self, offset: usize) -> Self::Slice { |
441 | #[cfg (debug_assertions)] |
442 | self.peek_slice(offset); |
443 | |
444 | // SAFETY: `Stream::next_slice_unchecked` requires `offset` to be in bounds |
445 | let slice = unsafe { self.get_unchecked(..offset) }; |
446 | slice |
447 | } |
448 | |
449 | #[inline (always)] |
450 | fn checkpoint(&self) -> Self::Checkpoint { |
451 | Checkpoint::<_, Self>::new(*self) |
452 | } |
453 | #[inline (always)] |
454 | fn reset(&mut self, checkpoint: &Self::Checkpoint) { |
455 | *self = checkpoint.inner; |
456 | } |
457 | |
458 | #[inline (always)] |
459 | fn raw(&self) -> &dyn crate::lib::std::fmt::Debug { |
460 | self |
461 | } |
462 | } |
463 | |
464 | impl<I> Stream for (I, usize) |
465 | where |
466 | I: Stream<Token = u8> + Clone, |
467 | { |
468 | type Token = bool; |
469 | type Slice = (I::Slice, usize, usize); |
470 | |
471 | type IterOffsets = BitOffsets<I>; |
472 | |
473 | type Checkpoint = Checkpoint<(I::Checkpoint, usize), Self>; |
474 | |
475 | #[inline (always)] |
476 | fn iter_offsets(&self) -> Self::IterOffsets { |
477 | BitOffsets { |
478 | i: self.clone(), |
479 | o: 0, |
480 | } |
481 | } |
482 | #[inline (always)] |
483 | fn eof_offset(&self) -> usize { |
484 | let offset = self.0.eof_offset() * 8; |
485 | if offset == 0 { |
486 | 0 |
487 | } else { |
488 | offset - self.1 |
489 | } |
490 | } |
491 | |
492 | #[inline (always)] |
493 | fn next_token(&mut self) -> Option<Self::Token> { |
494 | next_bit(self) |
495 | } |
496 | |
497 | #[inline (always)] |
498 | fn peek_token(&self) -> Option<Self::Token> { |
499 | peek_bit(self) |
500 | } |
501 | |
502 | #[inline (always)] |
503 | fn offset_for<P>(&self, predicate: P) -> Option<usize> |
504 | where |
505 | P: Fn(Self::Token) -> bool, |
506 | { |
507 | self.iter_offsets() |
508 | .find_map(|(o, b)| predicate(b).then_some(o)) |
509 | } |
510 | #[inline (always)] |
511 | fn offset_at(&self, tokens: usize) -> Result<usize, Needed> { |
512 | if let Some(needed) = tokens |
513 | .checked_sub(self.eof_offset()) |
514 | .and_then(NonZeroUsize::new) |
515 | { |
516 | Err(Needed::Size(needed)) |
517 | } else { |
518 | Ok(tokens) |
519 | } |
520 | } |
521 | #[inline (always)] |
522 | fn next_slice(&mut self, offset: usize) -> Self::Slice { |
523 | let byte_offset = (offset + self.1) / 8; |
524 | let end_offset = (offset + self.1) % 8; |
525 | let s = self.0.next_slice(byte_offset); |
526 | let start_offset = self.1; |
527 | self.1 = end_offset; |
528 | (s, start_offset, end_offset) |
529 | } |
530 | #[inline (always)] |
531 | fn peek_slice(&self, offset: usize) -> Self::Slice { |
532 | let byte_offset = (offset + self.1) / 8; |
533 | let end_offset = (offset + self.1) % 8; |
534 | let s = self.0.peek_slice(byte_offset); |
535 | let start_offset = self.1; |
536 | (s, start_offset, end_offset) |
537 | } |
538 | |
539 | #[inline (always)] |
540 | fn checkpoint(&self) -> Self::Checkpoint { |
541 | Checkpoint::<_, Self>::new((self.0.checkpoint(), self.1)) |
542 | } |
543 | #[inline (always)] |
544 | fn reset(&mut self, checkpoint: &Self::Checkpoint) { |
545 | self.0.reset(&checkpoint.inner.0); |
546 | self.1 = checkpoint.inner.1; |
547 | } |
548 | |
549 | #[inline (always)] |
550 | fn raw(&self) -> &dyn crate::lib::std::fmt::Debug { |
551 | &self.0 |
552 | } |
553 | } |
554 | |
555 | /// Iterator for [bit][crate::binary::bits] stream (`(I, usize)`) |
556 | pub struct BitOffsets<I> { |
557 | i: (I, usize), |
558 | o: usize, |
559 | } |
560 | |
561 | impl<I> Iterator for BitOffsets<I> |
562 | where |
563 | I: Stream<Token = u8> + Clone, |
564 | { |
565 | type Item = (usize, bool); |
566 | fn next(&mut self) -> Option<Self::Item> { |
567 | let b: bool = next_bit(&mut self.i)?; |
568 | let o: usize = self.o; |
569 | |
570 | self.o += 1; |
571 | |
572 | Some((o, b)) |
573 | } |
574 | } |
575 | |
576 | fn next_bit<I>(i: &mut (I, usize)) -> Option<bool> |
577 | where |
578 | I: Stream<Token = u8> + Clone, |
579 | { |
580 | if i.eof_offset() == 0 { |
581 | return None; |
582 | } |
583 | let offset: usize = i.1; |
584 | |
585 | let mut next_i: I = i.0.clone(); |
586 | let byte: u8 = next_i.next_token()?; |
587 | let bit: bool = (byte >> offset) & 0x1 == 0x1; |
588 | |
589 | let next_offset: usize = offset + 1; |
590 | if next_offset == 8 { |
591 | i.0 = next_i; |
592 | i.1 = 0; |
593 | Some(bit) |
594 | } else { |
595 | i.1 = next_offset; |
596 | Some(bit) |
597 | } |
598 | } |
599 | |
600 | fn peek_bit<I>(i: &(I, usize)) -> Option<bool> |
601 | where |
602 | I: Stream<Token = u8> + Clone, |
603 | { |
604 | if i.eof_offset() == 0 { |
605 | return None; |
606 | } |
607 | let offset: usize = i.1; |
608 | |
609 | let mut next_i: I = i.0.clone(); |
610 | let byte: u8 = next_i.next_token()?; |
611 | let bit: bool = (byte >> offset) & 0x1 == 0x1; |
612 | |
613 | let next_offset: usize = offset + 1; |
614 | if next_offset == 8 { |
615 | Some(bit) |
616 | } else { |
617 | Some(bit) |
618 | } |
619 | } |
620 | |
621 | /// Current parse locations offset |
622 | /// |
623 | /// See [`LocatingSlice`] for adding location tracking to your [`Stream`] |
624 | pub trait Location { |
625 | /// Previous token's end offset |
626 | fn previous_token_end(&self) -> usize; |
627 | /// Current token's start offset |
628 | fn current_token_start(&self) -> usize; |
629 | } |
630 | |
631 | /// Capture top-level errors in the middle of parsing so parsing can resume |
632 | /// |
633 | /// See [`Recoverable`] for adding error recovery tracking to your [`Stream`] |
634 | #[cfg (feature = "unstable-recover" )] |
635 | #[cfg (feature = "std" )] |
636 | pub trait Recover<E>: Stream { |
637 | /// Capture a top-level error |
638 | /// |
639 | /// May return `Err(err)` if recovery is not possible (e.g. if [`Recover::is_recovery_supported`] |
640 | /// returns `false`). |
641 | fn record_err( |
642 | &mut self, |
643 | token_start: &Self::Checkpoint, |
644 | err_start: &Self::Checkpoint, |
645 | err: E, |
646 | ) -> Result<(), E>; |
647 | |
648 | /// Report whether the [`Stream`] can save off errors for recovery |
649 | fn is_recovery_supported() -> bool; |
650 | } |
651 | |
652 | #[cfg (feature = "unstable-recover" )] |
653 | #[cfg (feature = "std" )] |
654 | impl<'a, T, E> Recover<E> for &'a [T] |
655 | where |
656 | &'a [T]: Stream, |
657 | { |
658 | #[inline (always)] |
659 | fn record_err( |
660 | &mut self, |
661 | _token_start: &Self::Checkpoint, |
662 | _err_start: &Self::Checkpoint, |
663 | err: E, |
664 | ) -> Result<(), E> { |
665 | Err(err) |
666 | } |
667 | |
668 | /// Report whether the [`Stream`] can save off errors for recovery |
669 | #[inline (always)] |
670 | fn is_recovery_supported() -> bool { |
671 | false |
672 | } |
673 | } |
674 | |
675 | #[cfg (feature = "unstable-recover" )] |
676 | #[cfg (feature = "std" )] |
677 | impl<E> Recover<E> for &str { |
678 | #[inline (always)] |
679 | fn record_err( |
680 | &mut self, |
681 | _token_start: &Self::Checkpoint, |
682 | _err_start: &Self::Checkpoint, |
683 | err: E, |
684 | ) -> Result<(), E> { |
685 | Err(err) |
686 | } |
687 | |
688 | /// Report whether the [`Stream`] can save off errors for recovery |
689 | #[inline (always)] |
690 | fn is_recovery_supported() -> bool { |
691 | false |
692 | } |
693 | } |
694 | |
695 | #[cfg (feature = "unstable-recover" )] |
696 | #[cfg (feature = "std" )] |
697 | impl<I, E> Recover<E> for (I, usize) |
698 | where |
699 | I: Recover<E>, |
700 | I: Stream<Token = u8> + Clone, |
701 | { |
702 | #[inline (always)] |
703 | fn record_err( |
704 | &mut self, |
705 | _token_start: &Self::Checkpoint, |
706 | _err_start: &Self::Checkpoint, |
707 | err: E, |
708 | ) -> Result<(), E> { |
709 | Err(err) |
710 | } |
711 | |
712 | /// Report whether the [`Stream`] can save off errors for recovery |
713 | #[inline (always)] |
714 | fn is_recovery_supported() -> bool { |
715 | false |
716 | } |
717 | } |
718 | |
719 | /// Marks the input as being the complete buffer or a partial buffer for streaming input |
720 | /// |
721 | /// See [`Partial`] for marking a presumed complete buffer type as a streaming buffer. |
722 | pub trait StreamIsPartial: Sized { |
723 | /// Whether the stream is currently partial or complete |
724 | type PartialState; |
725 | |
726 | /// Mark the stream is complete |
727 | #[must_use ] |
728 | fn complete(&mut self) -> Self::PartialState; |
729 | |
730 | /// Restore the stream back to its previous state |
731 | fn restore_partial(&mut self, state: Self::PartialState); |
732 | |
733 | /// Report whether the [`Stream`] is can ever be incomplete |
734 | fn is_partial_supported() -> bool; |
735 | |
736 | /// Report whether the [`Stream`] is currently incomplete |
737 | #[inline (always)] |
738 | fn is_partial(&self) -> bool { |
739 | Self::is_partial_supported() |
740 | } |
741 | } |
742 | |
743 | impl<T> StreamIsPartial for &[T] { |
744 | type PartialState = (); |
745 | |
746 | #[inline ] |
747 | fn complete(&mut self) -> Self::PartialState {} |
748 | |
749 | #[inline ] |
750 | fn restore_partial(&mut self, _state: Self::PartialState) {} |
751 | |
752 | #[inline (always)] |
753 | fn is_partial_supported() -> bool { |
754 | false |
755 | } |
756 | } |
757 | |
758 | impl StreamIsPartial for &str { |
759 | type PartialState = (); |
760 | |
761 | #[inline ] |
762 | fn complete(&mut self) -> Self::PartialState { |
763 | // Already complete |
764 | } |
765 | |
766 | #[inline ] |
767 | fn restore_partial(&mut self, _state: Self::PartialState) {} |
768 | |
769 | #[inline (always)] |
770 | fn is_partial_supported() -> bool { |
771 | false |
772 | } |
773 | } |
774 | |
775 | impl<I> StreamIsPartial for (I, usize) |
776 | where |
777 | I: StreamIsPartial, |
778 | { |
779 | type PartialState = I::PartialState; |
780 | |
781 | #[inline ] |
782 | fn complete(&mut self) -> Self::PartialState { |
783 | self.0.complete() |
784 | } |
785 | |
786 | #[inline ] |
787 | fn restore_partial(&mut self, state: Self::PartialState) { |
788 | self.0.restore_partial(state); |
789 | } |
790 | |
791 | #[inline (always)] |
792 | fn is_partial_supported() -> bool { |
793 | I::is_partial_supported() |
794 | } |
795 | |
796 | #[inline (always)] |
797 | fn is_partial(&self) -> bool { |
798 | self.0.is_partial() |
799 | } |
800 | } |
801 | |
802 | /// Useful functions to calculate the offset between slices and show a hexdump of a slice |
803 | pub trait Offset<Start = Self> { |
804 | /// Offset between the first byte of `start` and the first byte of `self`a |
805 | /// |
806 | /// <div class="warning"> |
807 | /// |
808 | /// **Note:** This is an offset, not an index, and may point to the end of input |
809 | /// (`start.len()`) when `self` is exhausted. |
810 | /// |
811 | /// </div> |
812 | fn offset_from(&self, start: &Start) -> usize; |
813 | } |
814 | |
815 | impl<T> Offset for &[T] { |
816 | #[inline ] |
817 | fn offset_from(&self, start: &Self) -> usize { |
818 | let fst: *const T = (*start).as_ptr(); |
819 | let snd: *const T = (*self).as_ptr(); |
820 | |
821 | debug_assert!( |
822 | fst <= snd, |
823 | "`Offset::offset_from( {snd:?}, {fst:?})` only accepts slices of `self`" |
824 | ); |
825 | (snd as usize - fst as usize) / crate::lib::std::mem::size_of::<T>() |
826 | } |
827 | } |
828 | |
829 | impl<'a, T> Offset<<&'a [T] as Stream>::Checkpoint> for &'a [T] |
830 | where |
831 | T: Clone + crate::lib::std::fmt::Debug, |
832 | { |
833 | #[inline (always)] |
834 | fn offset_from(&self, other: &<&'a [T] as Stream>::Checkpoint) -> usize { |
835 | self.checkpoint().offset_from(start:other) |
836 | } |
837 | } |
838 | |
839 | impl Offset for &str { |
840 | #[inline (always)] |
841 | fn offset_from(&self, start: &Self) -> usize { |
842 | self.as_bytes().offset_from(&start.as_bytes()) |
843 | } |
844 | } |
845 | |
846 | impl<'a> Offset<<&'a str as Stream>::Checkpoint> for &'a str { |
847 | #[inline (always)] |
848 | fn offset_from(&self, other: &<&'a str as Stream>::Checkpoint) -> usize { |
849 | self.checkpoint().offset_from(start:other) |
850 | } |
851 | } |
852 | |
853 | impl<I> Offset for (I, usize) |
854 | where |
855 | I: Offset, |
856 | { |
857 | #[inline (always)] |
858 | fn offset_from(&self, start: &Self) -> usize { |
859 | self.0.offset_from(&start.0) * 8 + self.1 - start.1 |
860 | } |
861 | } |
862 | |
863 | impl<I> Offset<<(I, usize) as Stream>::Checkpoint> for (I, usize) |
864 | where |
865 | I: Stream<Token = u8> + Clone, |
866 | { |
867 | #[inline (always)] |
868 | fn offset_from(&self, other: &<(I, usize) as Stream>::Checkpoint) -> usize { |
869 | self.checkpoint().offset_from(start:other) |
870 | } |
871 | } |
872 | |
873 | impl<I, S> Offset for Checkpoint<I, S> |
874 | where |
875 | I: Offset, |
876 | { |
877 | #[inline (always)] |
878 | fn offset_from(&self, start: &Self) -> usize { |
879 | self.inner.offset_from(&start.inner) |
880 | } |
881 | } |
882 | |
883 | /// Helper trait for types that can be viewed as a byte slice |
884 | pub trait AsBytes { |
885 | /// Casts the input type to a byte slice |
886 | fn as_bytes(&self) -> &[u8]; |
887 | } |
888 | |
889 | impl AsBytes for &[u8] { |
890 | #[inline (always)] |
891 | fn as_bytes(&self) -> &[u8] { |
892 | self |
893 | } |
894 | } |
895 | |
896 | /// Helper trait for types that can be viewed as a byte slice |
897 | pub trait AsBStr { |
898 | /// Casts the input type to a byte slice |
899 | fn as_bstr(&self) -> &[u8]; |
900 | } |
901 | |
902 | impl AsBStr for &[u8] { |
903 | #[inline (always)] |
904 | fn as_bstr(&self) -> &[u8] { |
905 | self |
906 | } |
907 | } |
908 | |
909 | impl AsBStr for &str { |
910 | #[inline (always)] |
911 | fn as_bstr(&self) -> &[u8] { |
912 | (*self).as_bytes() |
913 | } |
914 | } |
915 | |
916 | /// Result of [`Compare::compare`] |
917 | #[derive (Debug, Eq, PartialEq)] |
918 | pub enum CompareResult { |
919 | /// Comparison was successful |
920 | /// |
921 | /// `usize` is the end of the successful match within the buffer. |
922 | /// This is most relevant for caseless UTF-8 where `Compare::compare`'s parameter might be a different |
923 | /// length than the match within the buffer. |
924 | Ok(usize), |
925 | /// We need more data to be sure |
926 | Incomplete, |
927 | /// Comparison failed |
928 | Error, |
929 | } |
930 | |
931 | /// Abstracts comparison operations |
932 | pub trait Compare<T> { |
933 | /// Compares self to another value for equality |
934 | fn compare(&self, t: T) -> CompareResult; |
935 | } |
936 | |
937 | impl<'b> Compare<&'b [u8]> for &[u8] { |
938 | #[inline ] |
939 | fn compare(&self, t: &'b [u8]) -> CompareResult { |
940 | if t.iter().zip(*self).any(|(a: &u8, b: &u8)| a != b) { |
941 | CompareResult::Error |
942 | } else if self.len() < t.slice_len() { |
943 | CompareResult::Incomplete |
944 | } else { |
945 | CompareResult::Ok(t.slice_len()) |
946 | } |
947 | } |
948 | } |
949 | |
950 | impl<'b> Compare<AsciiCaseless<&'b [u8]>> for &[u8] { |
951 | #[inline ] |
952 | fn compare(&self, t: AsciiCaseless<&'b [u8]>) -> CompareResult { |
953 | if timpl Iterator .0 |
954 | .iter() |
955 | .zip(*self) |
956 | .any(|(a: &u8, b: &u8)| !a.eq_ignore_ascii_case(b)) |
957 | { |
958 | CompareResult::Error |
959 | } else if self.len() < t.slice_len() { |
960 | CompareResult::Incomplete |
961 | } else { |
962 | CompareResult::Ok(t.slice_len()) |
963 | } |
964 | } |
965 | } |
966 | |
967 | impl<const LEN: usize> Compare<[u8; LEN]> for &[u8] { |
968 | #[inline (always)] |
969 | fn compare(&self, t: [u8; LEN]) -> CompareResult { |
970 | self.compare(&t[..]) |
971 | } |
972 | } |
973 | |
974 | impl<const LEN: usize> Compare<AsciiCaseless<[u8; LEN]>> for &[u8] { |
975 | #[inline (always)] |
976 | fn compare(&self, t: AsciiCaseless<[u8; LEN]>) -> CompareResult { |
977 | self.compare(AsciiCaseless(&t.0[..])) |
978 | } |
979 | } |
980 | |
981 | impl<'b, const LEN: usize> Compare<&'b [u8; LEN]> for &[u8] { |
982 | #[inline (always)] |
983 | fn compare(&self, t: &'b [u8; LEN]) -> CompareResult { |
984 | self.compare(&t[..]) |
985 | } |
986 | } |
987 | |
988 | impl<'b, const LEN: usize> Compare<AsciiCaseless<&'b [u8; LEN]>> for &[u8] { |
989 | #[inline (always)] |
990 | fn compare(&self, t: AsciiCaseless<&'b [u8; LEN]>) -> CompareResult { |
991 | self.compare(AsciiCaseless(&t.0[..])) |
992 | } |
993 | } |
994 | |
995 | impl<'b> Compare<&'b str> for &[u8] { |
996 | #[inline (always)] |
997 | fn compare(&self, t: &'b str) -> CompareResult { |
998 | self.compare(t.as_bytes()) |
999 | } |
1000 | } |
1001 | |
1002 | impl<'b> Compare<AsciiCaseless<&'b str>> for &[u8] { |
1003 | #[inline (always)] |
1004 | fn compare(&self, t: AsciiCaseless<&'b str>) -> CompareResult { |
1005 | self.compare(AsciiCaseless(t.0.as_bytes())) |
1006 | } |
1007 | } |
1008 | |
1009 | impl Compare<u8> for &[u8] { |
1010 | #[inline ] |
1011 | fn compare(&self, t: u8) -> CompareResult { |
1012 | match self.first().copied() { |
1013 | Some(c: u8) if t == c => CompareResult::Ok(t.slice_len()), |
1014 | Some(_) => CompareResult::Error, |
1015 | None => CompareResult::Incomplete, |
1016 | } |
1017 | } |
1018 | } |
1019 | |
1020 | impl Compare<AsciiCaseless<u8>> for &[u8] { |
1021 | #[inline ] |
1022 | fn compare(&self, t: AsciiCaseless<u8>) -> CompareResult { |
1023 | match self.first() { |
1024 | Some(c: &u8) if t.0.eq_ignore_ascii_case(c) => CompareResult::Ok(t.slice_len()), |
1025 | Some(_) => CompareResult::Error, |
1026 | None => CompareResult::Incomplete, |
1027 | } |
1028 | } |
1029 | } |
1030 | |
1031 | impl Compare<char> for &[u8] { |
1032 | #[inline (always)] |
1033 | fn compare(&self, t: char) -> CompareResult { |
1034 | self.compare(t.encode_utf8(&mut [0; 4]).as_bytes()) |
1035 | } |
1036 | } |
1037 | |
1038 | impl Compare<AsciiCaseless<char>> for &[u8] { |
1039 | #[inline (always)] |
1040 | fn compare(&self, t: AsciiCaseless<char>) -> CompareResult { |
1041 | self.compare(AsciiCaseless(t.0.encode_utf8(&mut [0; 4]).as_bytes())) |
1042 | } |
1043 | } |
1044 | |
1045 | impl<'b> Compare<&'b str> for &str { |
1046 | #[inline (always)] |
1047 | fn compare(&self, t: &'b str) -> CompareResult { |
1048 | self.as_bytes().compare(t.as_bytes()) |
1049 | } |
1050 | } |
1051 | |
1052 | impl<'b> Compare<AsciiCaseless<&'b str>> for &str { |
1053 | #[inline (always)] |
1054 | fn compare(&self, t: AsciiCaseless<&'b str>) -> CompareResult { |
1055 | self.as_bytes().compare(t.as_bytes()) |
1056 | } |
1057 | } |
1058 | |
1059 | impl Compare<char> for &str { |
1060 | #[inline (always)] |
1061 | fn compare(&self, t: char) -> CompareResult { |
1062 | self.as_bytes().compare(t) |
1063 | } |
1064 | } |
1065 | |
1066 | impl Compare<AsciiCaseless<char>> for &str { |
1067 | #[inline (always)] |
1068 | fn compare(&self, t: AsciiCaseless<char>) -> CompareResult { |
1069 | self.as_bytes().compare(t) |
1070 | } |
1071 | } |
1072 | |
1073 | /// Look for a slice in self |
1074 | pub trait FindSlice<T> { |
1075 | /// Returns the offset of the slice if it is found |
1076 | fn find_slice(&self, substr: T) -> Option<crate::lib::std::ops::Range<usize>>; |
1077 | } |
1078 | |
1079 | impl<'s> FindSlice<&'s [u8]> for &[u8] { |
1080 | #[inline (always)] |
1081 | fn find_slice(&self, substr: &'s [u8]) -> Option<crate::lib::std::ops::Range<usize>> { |
1082 | memmem(self, literal:substr) |
1083 | } |
1084 | } |
1085 | |
1086 | impl<'s> FindSlice<(&'s [u8],)> for &[u8] { |
1087 | #[inline (always)] |
1088 | fn find_slice(&self, substr: (&'s [u8],)) -> Option<crate::lib::std::ops::Range<usize>> { |
1089 | memmem(self, literal:substr.0) |
1090 | } |
1091 | } |
1092 | |
1093 | impl<'s> FindSlice<(&'s [u8], &'s [u8])> for &[u8] { |
1094 | #[inline (always)] |
1095 | fn find_slice( |
1096 | &self, |
1097 | substr: (&'s [u8], &'s [u8]), |
1098 | ) -> Option<crate::lib::std::ops::Range<usize>> { |
1099 | memmem2(self, literal:substr) |
1100 | } |
1101 | } |
1102 | |
1103 | impl<'s> FindSlice<(&'s [u8], &'s [u8], &'s [u8])> for &[u8] { |
1104 | #[inline (always)] |
1105 | fn find_slice( |
1106 | &self, |
1107 | substr: (&'s [u8], &'s [u8], &'s [u8]), |
1108 | ) -> Option<crate::lib::std::ops::Range<usize>> { |
1109 | memmem3(self, literal:substr) |
1110 | } |
1111 | } |
1112 | |
1113 | impl FindSlice<char> for &[u8] { |
1114 | #[inline (always)] |
1115 | fn find_slice(&self, substr: char) -> Option<crate::lib::std::ops::Range<usize>> { |
1116 | let mut b: [u8; 4] = [0; 4]; |
1117 | let substr: &mut str = substr.encode_utf8(&mut b); |
1118 | self.find_slice(&*substr) |
1119 | } |
1120 | } |
1121 | |
1122 | impl FindSlice<(char,)> for &[u8] { |
1123 | #[inline (always)] |
1124 | fn find_slice(&self, substr: (char,)) -> Option<crate::lib::std::ops::Range<usize>> { |
1125 | let mut b: [u8; 4] = [0; 4]; |
1126 | let substr0: &mut str = substr.0.encode_utf8(&mut b); |
1127 | self.find_slice((&*substr0,)) |
1128 | } |
1129 | } |
1130 | |
1131 | impl FindSlice<(char, char)> for &[u8] { |
1132 | #[inline (always)] |
1133 | fn find_slice(&self, substr: (char, char)) -> Option<crate::lib::std::ops::Range<usize>> { |
1134 | let mut b: [u8; 4] = [0; 4]; |
1135 | let substr0: &mut str = substr.0.encode_utf8(&mut b); |
1136 | let mut b: [u8; 4] = [0; 4]; |
1137 | let substr1: &mut str = substr.1.encode_utf8(&mut b); |
1138 | self.find_slice((&*substr0, &*substr1)) |
1139 | } |
1140 | } |
1141 | |
1142 | impl FindSlice<(char, char, char)> for &[u8] { |
1143 | #[inline (always)] |
1144 | fn find_slice(&self, substr: (char, char, char)) -> Option<crate::lib::std::ops::Range<usize>> { |
1145 | let mut b: [u8; 4] = [0; 4]; |
1146 | let substr0: &mut str = substr.0.encode_utf8(&mut b); |
1147 | let mut b: [u8; 4] = [0; 4]; |
1148 | let substr1: &mut str = substr.1.encode_utf8(&mut b); |
1149 | let mut b: [u8; 4] = [0; 4]; |
1150 | let substr2: &mut str = substr.2.encode_utf8(&mut b); |
1151 | self.find_slice((&*substr0, &*substr1, &*substr2)) |
1152 | } |
1153 | } |
1154 | |
1155 | impl FindSlice<u8> for &[u8] { |
1156 | #[inline (always)] |
1157 | fn find_slice(&self, substr: u8) -> Option<crate::lib::std::ops::Range<usize>> { |
1158 | memchr(token:substr, self).map(|i: usize| i..i + 1) |
1159 | } |
1160 | } |
1161 | |
1162 | impl FindSlice<(u8,)> for &[u8] { |
1163 | #[inline (always)] |
1164 | fn find_slice(&self, substr: (u8,)) -> Option<crate::lib::std::ops::Range<usize>> { |
1165 | memchr(token:substr.0, self).map(|i: usize| i..i + 1) |
1166 | } |
1167 | } |
1168 | |
1169 | impl FindSlice<(u8, u8)> for &[u8] { |
1170 | #[inline (always)] |
1171 | fn find_slice(&self, substr: (u8, u8)) -> Option<crate::lib::std::ops::Range<usize>> { |
1172 | memchr2(token:substr, self).map(|i: usize| i..i + 1) |
1173 | } |
1174 | } |
1175 | |
1176 | impl FindSlice<(u8, u8, u8)> for &[u8] { |
1177 | #[inline (always)] |
1178 | fn find_slice(&self, substr: (u8, u8, u8)) -> Option<crate::lib::std::ops::Range<usize>> { |
1179 | memchr3(token:substr, self).map(|i: usize| i..i + 1) |
1180 | } |
1181 | } |
1182 | |
1183 | impl<'s> FindSlice<&'s str> for &[u8] { |
1184 | #[inline (always)] |
1185 | fn find_slice(&self, substr: &'s str) -> Option<crate::lib::std::ops::Range<usize>> { |
1186 | self.find_slice(substr.as_bytes()) |
1187 | } |
1188 | } |
1189 | |
1190 | impl<'s> FindSlice<(&'s str,)> for &[u8] { |
1191 | #[inline (always)] |
1192 | fn find_slice(&self, substr: (&'s str,)) -> Option<crate::lib::std::ops::Range<usize>> { |
1193 | memmem(self, literal:substr.0.as_bytes()) |
1194 | } |
1195 | } |
1196 | |
1197 | impl<'s> FindSlice<(&'s str, &'s str)> for &[u8] { |
1198 | #[inline (always)] |
1199 | fn find_slice(&self, substr: (&'s str, &'s str)) -> Option<crate::lib::std::ops::Range<usize>> { |
1200 | memmem2(self, (substr.0.as_bytes(), substr.1.as_bytes())) |
1201 | } |
1202 | } |
1203 | |
1204 | impl<'s> FindSlice<(&'s str, &'s str, &'s str)> for &[u8] { |
1205 | #[inline (always)] |
1206 | fn find_slice( |
1207 | &self, |
1208 | substr: (&'s str, &'s str, &'s str), |
1209 | ) -> Option<crate::lib::std::ops::Range<usize>> { |
1210 | memmem3( |
1211 | self, |
1212 | ( |
1213 | substr.0.as_bytes(), |
1214 | substr.1.as_bytes(), |
1215 | substr.2.as_bytes(), |
1216 | ), |
1217 | ) |
1218 | } |
1219 | } |
1220 | |
1221 | impl<'s> FindSlice<&'s str> for &str { |
1222 | #[inline (always)] |
1223 | fn find_slice(&self, substr: &'s str) -> Option<crate::lib::std::ops::Range<usize>> { |
1224 | self.as_bytes().find_slice(substr) |
1225 | } |
1226 | } |
1227 | |
1228 | impl<'s> FindSlice<(&'s str,)> for &str { |
1229 | #[inline (always)] |
1230 | fn find_slice(&self, substr: (&'s str,)) -> Option<crate::lib::std::ops::Range<usize>> { |
1231 | self.as_bytes().find_slice(substr) |
1232 | } |
1233 | } |
1234 | |
1235 | impl<'s> FindSlice<(&'s str, &'s str)> for &str { |
1236 | #[inline (always)] |
1237 | fn find_slice(&self, substr: (&'s str, &'s str)) -> Option<crate::lib::std::ops::Range<usize>> { |
1238 | self.as_bytes().find_slice(substr) |
1239 | } |
1240 | } |
1241 | |
1242 | impl<'s> FindSlice<(&'s str, &'s str, &'s str)> for &str { |
1243 | #[inline (always)] |
1244 | fn find_slice( |
1245 | &self, |
1246 | substr: (&'s str, &'s str, &'s str), |
1247 | ) -> Option<crate::lib::std::ops::Range<usize>> { |
1248 | self.as_bytes().find_slice(substr) |
1249 | } |
1250 | } |
1251 | |
1252 | impl FindSlice<char> for &str { |
1253 | #[inline (always)] |
1254 | fn find_slice(&self, substr: char) -> Option<crate::lib::std::ops::Range<usize>> { |
1255 | self.as_bytes().find_slice(substr) |
1256 | } |
1257 | } |
1258 | |
1259 | impl FindSlice<(char,)> for &str { |
1260 | #[inline (always)] |
1261 | fn find_slice(&self, substr: (char,)) -> Option<crate::lib::std::ops::Range<usize>> { |
1262 | self.as_bytes().find_slice(substr) |
1263 | } |
1264 | } |
1265 | |
1266 | impl FindSlice<(char, char)> for &str { |
1267 | #[inline (always)] |
1268 | fn find_slice(&self, substr: (char, char)) -> Option<crate::lib::std::ops::Range<usize>> { |
1269 | self.as_bytes().find_slice(substr) |
1270 | } |
1271 | } |
1272 | |
1273 | impl FindSlice<(char, char, char)> for &str { |
1274 | #[inline (always)] |
1275 | fn find_slice(&self, substr: (char, char, char)) -> Option<crate::lib::std::ops::Range<usize>> { |
1276 | self.as_bytes().find_slice(substr) |
1277 | } |
1278 | } |
1279 | |
1280 | /// Used to integrate `str`'s `parse()` method |
1281 | pub trait ParseSlice<R> { |
1282 | /// Succeeds if `parse()` succeeded |
1283 | /// |
1284 | /// The byte slice implementation will first convert it to a `&str`, then apply the `parse()` |
1285 | /// function |
1286 | fn parse_slice(&self) -> Option<R>; |
1287 | } |
1288 | |
1289 | impl<R: FromStr> ParseSlice<R> for &[u8] { |
1290 | #[inline (always)] |
1291 | fn parse_slice(&self) -> Option<R> { |
1292 | from_utf8(self).ok().and_then(|s: &str| s.parse().ok()) |
1293 | } |
1294 | } |
1295 | |
1296 | impl<R: FromStr> ParseSlice<R> for &str { |
1297 | #[inline (always)] |
1298 | fn parse_slice(&self) -> Option<R> { |
1299 | self.parse().ok() |
1300 | } |
1301 | } |
1302 | |
1303 | /// Convert a `Stream` into an appropriate `Output` type |
1304 | pub trait UpdateSlice: Stream { |
1305 | /// Convert an `Output` type to be used as `Stream` |
1306 | fn update_slice(self, inner: Self::Slice) -> Self; |
1307 | } |
1308 | |
1309 | impl<T> UpdateSlice for &[T] |
1310 | where |
1311 | T: Clone + crate::lib::std::fmt::Debug, |
1312 | { |
1313 | #[inline (always)] |
1314 | fn update_slice(self, inner: Self::Slice) -> Self { |
1315 | inner |
1316 | } |
1317 | } |
1318 | |
1319 | impl UpdateSlice for &str { |
1320 | #[inline (always)] |
1321 | fn update_slice(self, inner: Self::Slice) -> Self { |
1322 | inner |
1323 | } |
1324 | } |
1325 | |
1326 | /// Ensure checkpoint details are kept private |
1327 | pub struct Checkpoint<T, S> { |
1328 | inner: T, |
1329 | stream: core::marker::PhantomData<S>, |
1330 | } |
1331 | |
1332 | impl<T, S> Checkpoint<T, S> { |
1333 | fn new(inner: T) -> Self { |
1334 | Self { |
1335 | inner, |
1336 | stream: Default::default(), |
1337 | } |
1338 | } |
1339 | } |
1340 | |
1341 | impl<T: Copy, S> Copy for Checkpoint<T, S> {} |
1342 | |
1343 | impl<T: Clone, S> Clone for Checkpoint<T, S> { |
1344 | #[inline (always)] |
1345 | fn clone(&self) -> Self { |
1346 | Self { |
1347 | inner: self.inner.clone(), |
1348 | stream: Default::default(), |
1349 | } |
1350 | } |
1351 | } |
1352 | |
1353 | impl<T: PartialOrd, S> PartialOrd for Checkpoint<T, S> { |
1354 | #[inline (always)] |
1355 | fn partial_cmp(&self, other: &Self) -> Option<core::cmp::Ordering> { |
1356 | self.inner.partial_cmp(&other.inner) |
1357 | } |
1358 | } |
1359 | |
1360 | impl<T: Ord, S> Ord for Checkpoint<T, S> { |
1361 | #[inline (always)] |
1362 | fn cmp(&self, other: &Self) -> core::cmp::Ordering { |
1363 | self.inner.cmp(&other.inner) |
1364 | } |
1365 | } |
1366 | |
1367 | impl<T: PartialEq, S> PartialEq for Checkpoint<T, S> { |
1368 | #[inline (always)] |
1369 | fn eq(&self, other: &Self) -> bool { |
1370 | self.inner.eq(&other.inner) |
1371 | } |
1372 | } |
1373 | |
1374 | impl<T: Eq, S> Eq for Checkpoint<T, S> {} |
1375 | |
1376 | impl<T: crate::lib::std::fmt::Debug, S> crate::lib::std::fmt::Debug for Checkpoint<T, S> { |
1377 | fn fmt(&self, f: &mut crate::lib::std::fmt::Formatter<'_>) -> crate::lib::std::fmt::Result { |
1378 | self.inner.fmt(f) |
1379 | } |
1380 | } |
1381 | |
1382 | /// Abstracts something which can extend an `Extend`. |
1383 | /// Used to build modified input slices in `escaped_transform` |
1384 | pub trait Accumulate<T>: Sized { |
1385 | /// Create a new `Extend` of the correct type |
1386 | fn initial(capacity: Option<usize>) -> Self; |
1387 | /// Accumulate the input into an accumulator |
1388 | fn accumulate(&mut self, acc: T); |
1389 | } |
1390 | |
1391 | impl<T> Accumulate<T> for () { |
1392 | #[inline (always)] |
1393 | fn initial(_capacity: Option<usize>) -> Self {} |
1394 | #[inline (always)] |
1395 | fn accumulate(&mut self, _acc: T) {} |
1396 | } |
1397 | |
1398 | impl<T> Accumulate<T> for usize { |
1399 | #[inline (always)] |
1400 | fn initial(_capacity: Option<usize>) -> Self { |
1401 | 0 |
1402 | } |
1403 | #[inline (always)] |
1404 | fn accumulate(&mut self, _acc: T) { |
1405 | *self += 1; |
1406 | } |
1407 | } |
1408 | |
1409 | #[cfg (feature = "alloc" )] |
1410 | impl<T> Accumulate<T> for Vec<T> { |
1411 | #[inline (always)] |
1412 | fn initial(capacity: Option<usize>) -> Self { |
1413 | match capacity { |
1414 | Some(capacity: usize) => Vec::with_capacity(clamp_capacity::<T>(capacity)), |
1415 | None => Vec::new(), |
1416 | } |
1417 | } |
1418 | #[inline (always)] |
1419 | fn accumulate(&mut self, acc: T) { |
1420 | self.push(acc); |
1421 | } |
1422 | } |
1423 | |
1424 | #[cfg (feature = "alloc" )] |
1425 | impl<'i, T: Clone> Accumulate<&'i [T]> for Vec<T> { |
1426 | #[inline (always)] |
1427 | fn initial(capacity: Option<usize>) -> Self { |
1428 | match capacity { |
1429 | Some(capacity: usize) => Vec::with_capacity(clamp_capacity::<T>(capacity)), |
1430 | None => Vec::new(), |
1431 | } |
1432 | } |
1433 | #[inline (always)] |
1434 | fn accumulate(&mut self, acc: &'i [T]) { |
1435 | self.extend(iter:acc.iter().cloned()); |
1436 | } |
1437 | } |
1438 | |
1439 | #[cfg (feature = "alloc" )] |
1440 | impl Accumulate<char> for String { |
1441 | #[inline (always)] |
1442 | fn initial(capacity: Option<usize>) -> Self { |
1443 | match capacity { |
1444 | Some(capacity: usize) => String::with_capacity(clamp_capacity::<char>(capacity)), |
1445 | None => String::new(), |
1446 | } |
1447 | } |
1448 | #[inline (always)] |
1449 | fn accumulate(&mut self, acc: char) { |
1450 | self.push(ch:acc); |
1451 | } |
1452 | } |
1453 | |
1454 | #[cfg (feature = "alloc" )] |
1455 | impl<'i> Accumulate<&'i str> for String { |
1456 | #[inline (always)] |
1457 | fn initial(capacity: Option<usize>) -> Self { |
1458 | match capacity { |
1459 | Some(capacity: usize) => String::with_capacity(clamp_capacity::<char>(capacity)), |
1460 | None => String::new(), |
1461 | } |
1462 | } |
1463 | #[inline (always)] |
1464 | fn accumulate(&mut self, acc: &'i str) { |
1465 | self.push_str(string:acc); |
1466 | } |
1467 | } |
1468 | |
1469 | #[cfg (feature = "alloc" )] |
1470 | impl<K, V> Accumulate<(K, V)> for BTreeMap<K, V> |
1471 | where |
1472 | K: crate::lib::std::cmp::Ord, |
1473 | { |
1474 | #[inline (always)] |
1475 | fn initial(_capacity: Option<usize>) -> Self { |
1476 | BTreeMap::new() |
1477 | } |
1478 | #[inline (always)] |
1479 | fn accumulate(&mut self, (key: K, value: V): (K, V)) { |
1480 | self.insert(key, value); |
1481 | } |
1482 | } |
1483 | |
1484 | #[cfg (feature = "std" )] |
1485 | impl<K, V, S> Accumulate<(K, V)> for HashMap<K, V, S> |
1486 | where |
1487 | K: crate::lib::std::cmp::Eq + crate::lib::std::hash::Hash, |
1488 | S: BuildHasher + Default, |
1489 | { |
1490 | #[inline (always)] |
1491 | fn initial(capacity: Option<usize>) -> Self { |
1492 | let h: S = S::default(); |
1493 | match capacity { |
1494 | Some(capacity: usize) => { |
1495 | HashMap::with_capacity_and_hasher(clamp_capacity::<(K, V)>(capacity), hasher:h) |
1496 | } |
1497 | None => HashMap::with_hasher(hash_builder:h), |
1498 | } |
1499 | } |
1500 | #[inline (always)] |
1501 | fn accumulate(&mut self, (key: K, value: V): (K, V)) { |
1502 | self.insert(k:key, v:value); |
1503 | } |
1504 | } |
1505 | |
1506 | #[cfg (feature = "alloc" )] |
1507 | impl<K> Accumulate<K> for BTreeSet<K> |
1508 | where |
1509 | K: crate::lib::std::cmp::Ord, |
1510 | { |
1511 | #[inline (always)] |
1512 | fn initial(_capacity: Option<usize>) -> Self { |
1513 | BTreeSet::new() |
1514 | } |
1515 | #[inline (always)] |
1516 | fn accumulate(&mut self, key: K) { |
1517 | self.insert(key); |
1518 | } |
1519 | } |
1520 | |
1521 | #[cfg (feature = "std" )] |
1522 | impl<K, S> Accumulate<K> for HashSet<K, S> |
1523 | where |
1524 | K: crate::lib::std::cmp::Eq + crate::lib::std::hash::Hash, |
1525 | S: BuildHasher + Default, |
1526 | { |
1527 | #[inline (always)] |
1528 | fn initial(capacity: Option<usize>) -> Self { |
1529 | let h: S = S::default(); |
1530 | match capacity { |
1531 | Some(capacity: usize) => HashSet::with_capacity_and_hasher(clamp_capacity::<K>(capacity), hasher:h), |
1532 | None => HashSet::with_hasher(h), |
1533 | } |
1534 | } |
1535 | #[inline (always)] |
1536 | fn accumulate(&mut self, key: K) { |
1537 | self.insert(key); |
1538 | } |
1539 | } |
1540 | |
1541 | #[cfg (feature = "alloc" )] |
1542 | #[inline ] |
1543 | pub(crate) fn clamp_capacity<T>(capacity: usize) -> usize { |
1544 | /// Don't pre-allocate more than 64KiB when calling `Vec::with_capacity`. |
1545 | /// |
1546 | /// Pre-allocating memory is a nice optimization but count fields can't |
1547 | /// always be trusted. We should clamp initial capacities to some reasonable |
1548 | /// amount. This reduces the risk of a bogus count value triggering a panic |
1549 | /// due to an OOM error. |
1550 | /// |
1551 | /// This does not affect correctness. `winnow` will always read the full number |
1552 | /// of elements regardless of the capacity cap. |
1553 | const MAX_INITIAL_CAPACITY_BYTES: usize = 65536; |
1554 | |
1555 | let max_initial_capacity: usize = |
1556 | MAX_INITIAL_CAPACITY_BYTES / crate::lib::std::mem::size_of::<T>().max(1); |
1557 | capacity.min(max_initial_capacity) |
1558 | } |
1559 | |
1560 | /// Helper trait to convert numbers to usize. |
1561 | /// |
1562 | /// By default, usize implements `From<u8>` and `From<u16>` but not |
1563 | /// `From<u32>` and `From<u64>` because that would be invalid on some |
1564 | /// platforms. This trait implements the conversion for platforms |
1565 | /// with 32 and 64 bits pointer platforms |
1566 | pub trait ToUsize { |
1567 | /// converts self to usize |
1568 | fn to_usize(&self) -> usize; |
1569 | } |
1570 | |
1571 | impl ToUsize for u8 { |
1572 | #[inline (always)] |
1573 | fn to_usize(&self) -> usize { |
1574 | *self as usize |
1575 | } |
1576 | } |
1577 | |
1578 | impl ToUsize for u16 { |
1579 | #[inline (always)] |
1580 | fn to_usize(&self) -> usize { |
1581 | *self as usize |
1582 | } |
1583 | } |
1584 | |
1585 | impl ToUsize for usize { |
1586 | #[inline (always)] |
1587 | fn to_usize(&self) -> usize { |
1588 | *self |
1589 | } |
1590 | } |
1591 | |
1592 | #[cfg (any(target_pointer_width = "32" , target_pointer_width = "64" ))] |
1593 | impl ToUsize for u32 { |
1594 | #[inline (always)] |
1595 | fn to_usize(&self) -> usize { |
1596 | *self as usize |
1597 | } |
1598 | } |
1599 | |
1600 | #[cfg (target_pointer_width = "64" )] |
1601 | impl ToUsize for u64 { |
1602 | #[inline (always)] |
1603 | fn to_usize(&self) -> usize { |
1604 | *self as usize |
1605 | } |
1606 | } |
1607 | |
1608 | /// Transforms a token into a char for basic string parsing |
1609 | #[allow (clippy::len_without_is_empty)] |
1610 | #[allow (clippy::wrong_self_convention)] |
1611 | pub trait AsChar { |
1612 | /// Makes a char from self |
1613 | /// |
1614 | /// # Example |
1615 | /// |
1616 | /// ``` |
1617 | /// use winnow::prelude::*; |
1618 | /// |
1619 | /// assert_eq!('a' .as_char(), 'a' ); |
1620 | /// assert_eq!(u8::MAX.as_char(), std::char::from_u32(u8::MAX as u32).unwrap()); |
1621 | /// ``` |
1622 | fn as_char(self) -> char; |
1623 | |
1624 | /// Tests that self is an alphabetic character |
1625 | /// |
1626 | /// <div class="warning"> |
1627 | /// |
1628 | /// **Warning:** for `&str` it matches alphabetic |
1629 | /// characters outside of the 52 ASCII letters |
1630 | /// |
1631 | /// </div> |
1632 | fn is_alpha(self) -> bool; |
1633 | |
1634 | /// Tests that self is an alphabetic character |
1635 | /// or a decimal digit |
1636 | fn is_alphanum(self) -> bool; |
1637 | /// Tests that self is a decimal digit |
1638 | fn is_dec_digit(self) -> bool; |
1639 | /// Tests that self is an hex digit |
1640 | fn is_hex_digit(self) -> bool; |
1641 | /// Tests that self is an octal digit |
1642 | fn is_oct_digit(self) -> bool; |
1643 | /// Gets the len in bytes for self |
1644 | fn len(self) -> usize; |
1645 | /// Tests that self is ASCII space or tab |
1646 | fn is_space(self) -> bool; |
1647 | /// Tests if byte is ASCII newline: \n |
1648 | fn is_newline(self) -> bool; |
1649 | } |
1650 | |
1651 | impl AsChar for u8 { |
1652 | #[inline (always)] |
1653 | fn as_char(self) -> char { |
1654 | self as char |
1655 | } |
1656 | #[inline ] |
1657 | fn is_alpha(self) -> bool { |
1658 | matches!(self, 0x41..=0x5A | 0x61..=0x7A) |
1659 | } |
1660 | #[inline ] |
1661 | fn is_alphanum(self) -> bool { |
1662 | self.is_alpha() || self.is_dec_digit() |
1663 | } |
1664 | #[inline ] |
1665 | fn is_dec_digit(self) -> bool { |
1666 | matches!(self, 0x30..=0x39) |
1667 | } |
1668 | #[inline ] |
1669 | fn is_hex_digit(self) -> bool { |
1670 | matches!(self, 0x30..=0x39 | 0x41..=0x46 | 0x61..=0x66) |
1671 | } |
1672 | #[inline ] |
1673 | fn is_oct_digit(self) -> bool { |
1674 | matches!(self, 0x30..=0x37) |
1675 | } |
1676 | #[inline ] |
1677 | fn len(self) -> usize { |
1678 | 1 |
1679 | } |
1680 | #[inline ] |
1681 | fn is_space(self) -> bool { |
1682 | self == b' ' || self == b' \t' |
1683 | } |
1684 | #[inline ] |
1685 | fn is_newline(self) -> bool { |
1686 | self == b' \n' |
1687 | } |
1688 | } |
1689 | |
1690 | impl AsChar for &u8 { |
1691 | #[inline (always)] |
1692 | fn as_char(self) -> char { |
1693 | (*self).as_char() |
1694 | } |
1695 | #[inline (always)] |
1696 | fn is_alpha(self) -> bool { |
1697 | (*self).is_alpha() |
1698 | } |
1699 | #[inline (always)] |
1700 | fn is_alphanum(self) -> bool { |
1701 | (*self).is_alphanum() |
1702 | } |
1703 | #[inline (always)] |
1704 | fn is_dec_digit(self) -> bool { |
1705 | (*self).is_dec_digit() |
1706 | } |
1707 | #[inline (always)] |
1708 | fn is_hex_digit(self) -> bool { |
1709 | (*self).is_hex_digit() |
1710 | } |
1711 | #[inline (always)] |
1712 | fn is_oct_digit(self) -> bool { |
1713 | (*self).is_oct_digit() |
1714 | } |
1715 | #[inline (always)] |
1716 | fn len(self) -> usize { |
1717 | (*self).len() |
1718 | } |
1719 | #[inline (always)] |
1720 | fn is_space(self) -> bool { |
1721 | (*self).is_space() |
1722 | } |
1723 | #[inline (always)] |
1724 | fn is_newline(self) -> bool { |
1725 | (*self).is_newline() |
1726 | } |
1727 | } |
1728 | |
1729 | impl AsChar for char { |
1730 | #[inline (always)] |
1731 | fn as_char(self) -> char { |
1732 | self |
1733 | } |
1734 | #[inline ] |
1735 | fn is_alpha(self) -> bool { |
1736 | self.is_ascii_alphabetic() |
1737 | } |
1738 | #[inline ] |
1739 | fn is_alphanum(self) -> bool { |
1740 | self.is_alpha() || self.is_dec_digit() |
1741 | } |
1742 | #[inline ] |
1743 | fn is_dec_digit(self) -> bool { |
1744 | self.is_ascii_digit() |
1745 | } |
1746 | #[inline ] |
1747 | fn is_hex_digit(self) -> bool { |
1748 | self.is_ascii_hexdigit() |
1749 | } |
1750 | #[inline ] |
1751 | fn is_oct_digit(self) -> bool { |
1752 | self.is_digit(8) |
1753 | } |
1754 | #[inline ] |
1755 | fn len(self) -> usize { |
1756 | self.len_utf8() |
1757 | } |
1758 | #[inline ] |
1759 | fn is_space(self) -> bool { |
1760 | self == ' ' || self == ' \t' |
1761 | } |
1762 | #[inline ] |
1763 | fn is_newline(self) -> bool { |
1764 | self == ' \n' |
1765 | } |
1766 | } |
1767 | |
1768 | impl AsChar for &char { |
1769 | #[inline (always)] |
1770 | fn as_char(self) -> char { |
1771 | (*self).as_char() |
1772 | } |
1773 | #[inline (always)] |
1774 | fn is_alpha(self) -> bool { |
1775 | (*self).is_alpha() |
1776 | } |
1777 | #[inline (always)] |
1778 | fn is_alphanum(self) -> bool { |
1779 | (*self).is_alphanum() |
1780 | } |
1781 | #[inline (always)] |
1782 | fn is_dec_digit(self) -> bool { |
1783 | (*self).is_dec_digit() |
1784 | } |
1785 | #[inline (always)] |
1786 | fn is_hex_digit(self) -> bool { |
1787 | (*self).is_hex_digit() |
1788 | } |
1789 | #[inline (always)] |
1790 | fn is_oct_digit(self) -> bool { |
1791 | (*self).is_oct_digit() |
1792 | } |
1793 | #[inline (always)] |
1794 | fn len(self) -> usize { |
1795 | (*self).len() |
1796 | } |
1797 | #[inline (always)] |
1798 | fn is_space(self) -> bool { |
1799 | (*self).is_space() |
1800 | } |
1801 | #[inline (always)] |
1802 | fn is_newline(self) -> bool { |
1803 | (*self).is_newline() |
1804 | } |
1805 | } |
1806 | |
1807 | /// Check if a token is in a set of possible tokens |
1808 | /// |
1809 | /// While this can be implemented manually, you can also build up sets using: |
1810 | /// - `b'c'` and `'c'` |
1811 | /// - `b""` |
1812 | /// - `|c| true` |
1813 | /// - `b'a'..=b'z'`, `'a'..='z'` (etc for each [range type][std::ops]) |
1814 | /// - `(set1, set2, ...)` |
1815 | /// |
1816 | /// # Example |
1817 | /// |
1818 | /// For example, you could implement `hex_digit0` as: |
1819 | /// ``` |
1820 | /// # use winnow::prelude::*; |
1821 | /// # use winnow::{error::ErrMode, error::ContextError}; |
1822 | /// # use winnow::token::take_while; |
1823 | /// fn hex_digit1<'s>(input: &mut &'s str) -> ModalResult<&'s str, ContextError> { |
1824 | /// take_while(1.., ('a' ..='f' , 'A' ..='F' , '0' ..='9' )).parse_next(input) |
1825 | /// } |
1826 | /// |
1827 | /// assert_eq!(hex_digit1.parse_peek("21cZ" ), Ok(("Z" , "21c" ))); |
1828 | /// assert!(hex_digit1.parse_peek("H2" ).is_err()); |
1829 | /// assert!(hex_digit1.parse_peek("" ).is_err()); |
1830 | /// ``` |
1831 | pub trait ContainsToken<T> { |
1832 | /// Returns true if self contains the token |
1833 | fn contains_token(&self, token: T) -> bool; |
1834 | } |
1835 | |
1836 | impl ContainsToken<u8> for u8 { |
1837 | #[inline (always)] |
1838 | fn contains_token(&self, token: u8) -> bool { |
1839 | *self == token |
1840 | } |
1841 | } |
1842 | |
1843 | impl ContainsToken<&u8> for u8 { |
1844 | #[inline (always)] |
1845 | fn contains_token(&self, token: &u8) -> bool { |
1846 | self.contains_token(*token) |
1847 | } |
1848 | } |
1849 | |
1850 | impl ContainsToken<char> for u8 { |
1851 | #[inline (always)] |
1852 | fn contains_token(&self, token: char) -> bool { |
1853 | self.as_char() == token |
1854 | } |
1855 | } |
1856 | |
1857 | impl ContainsToken<&char> for u8 { |
1858 | #[inline (always)] |
1859 | fn contains_token(&self, token: &char) -> bool { |
1860 | self.contains_token(*token) |
1861 | } |
1862 | } |
1863 | |
1864 | impl<C: AsChar> ContainsToken<C> for char { |
1865 | #[inline (always)] |
1866 | fn contains_token(&self, token: C) -> bool { |
1867 | *self == token.as_char() |
1868 | } |
1869 | } |
1870 | |
1871 | impl<C, F: Fn(C) -> bool> ContainsToken<C> for F { |
1872 | #[inline (always)] |
1873 | fn contains_token(&self, token: C) -> bool { |
1874 | self(token) |
1875 | } |
1876 | } |
1877 | |
1878 | impl<C1: AsChar, C2: AsChar + Clone> ContainsToken<C1> for crate::lib::std::ops::Range<C2> { |
1879 | #[inline (always)] |
1880 | fn contains_token(&self, token: C1) -> bool { |
1881 | let start: char = self.start.clone().as_char(); |
1882 | let end: char = self.end.clone().as_char(); |
1883 | (start..end).contains(&token.as_char()) |
1884 | } |
1885 | } |
1886 | |
1887 | impl<C1: AsChar, C2: AsChar + Clone> ContainsToken<C1> |
1888 | for crate::lib::std::ops::RangeInclusive<C2> |
1889 | { |
1890 | #[inline (always)] |
1891 | fn contains_token(&self, token: C1) -> bool { |
1892 | let start: char = self.start().clone().as_char(); |
1893 | let end: char = self.end().clone().as_char(); |
1894 | (start..=end).contains(&token.as_char()) |
1895 | } |
1896 | } |
1897 | |
1898 | impl<C1: AsChar, C2: AsChar + Clone> ContainsToken<C1> for crate::lib::std::ops::RangeFrom<C2> { |
1899 | #[inline (always)] |
1900 | fn contains_token(&self, token: C1) -> bool { |
1901 | let start: char = self.start.clone().as_char(); |
1902 | (start..).contains(&token.as_char()) |
1903 | } |
1904 | } |
1905 | |
1906 | impl<C1: AsChar, C2: AsChar + Clone> ContainsToken<C1> for crate::lib::std::ops::RangeTo<C2> { |
1907 | #[inline (always)] |
1908 | fn contains_token(&self, token: C1) -> bool { |
1909 | let end: char = self.end.clone().as_char(); |
1910 | (..end).contains(&token.as_char()) |
1911 | } |
1912 | } |
1913 | |
1914 | impl<C1: AsChar, C2: AsChar + Clone> ContainsToken<C1> |
1915 | for crate::lib::std::ops::RangeToInclusive<C2> |
1916 | { |
1917 | #[inline (always)] |
1918 | fn contains_token(&self, token: C1) -> bool { |
1919 | let end: char = self.end.clone().as_char(); |
1920 | (..=end).contains(&token.as_char()) |
1921 | } |
1922 | } |
1923 | |
1924 | impl<C1: AsChar> ContainsToken<C1> for crate::lib::std::ops::RangeFull { |
1925 | #[inline (always)] |
1926 | fn contains_token(&self, _token: C1) -> bool { |
1927 | true |
1928 | } |
1929 | } |
1930 | |
1931 | impl<C: AsChar> ContainsToken<C> for &'_ [u8] { |
1932 | #[inline ] |
1933 | fn contains_token(&self, token: C) -> bool { |
1934 | let token: char = token.as_char(); |
1935 | self.iter().any(|t: &u8| t.as_char() == token) |
1936 | } |
1937 | } |
1938 | |
1939 | impl<C: AsChar> ContainsToken<C> for &'_ [char] { |
1940 | #[inline ] |
1941 | fn contains_token(&self, token: C) -> bool { |
1942 | let token: char = token.as_char(); |
1943 | self.iter().any(|t: &char| *t == token) |
1944 | } |
1945 | } |
1946 | |
1947 | impl<const LEN: usize, C: AsChar> ContainsToken<C> for &'_ [u8; LEN] { |
1948 | #[inline ] |
1949 | fn contains_token(&self, token: C) -> bool { |
1950 | let token: char = token.as_char(); |
1951 | self.iter().any(|t: &u8| t.as_char() == token) |
1952 | } |
1953 | } |
1954 | |
1955 | impl<const LEN: usize, C: AsChar> ContainsToken<C> for &'_ [char; LEN] { |
1956 | #[inline ] |
1957 | fn contains_token(&self, token: C) -> bool { |
1958 | let token: char = token.as_char(); |
1959 | self.iter().any(|t: &char| *t == token) |
1960 | } |
1961 | } |
1962 | |
1963 | impl<const LEN: usize, C: AsChar> ContainsToken<C> for [u8; LEN] { |
1964 | #[inline ] |
1965 | fn contains_token(&self, token: C) -> bool { |
1966 | let token: char = token.as_char(); |
1967 | self.iter().any(|t: &u8| t.as_char() == token) |
1968 | } |
1969 | } |
1970 | |
1971 | impl<const LEN: usize, C: AsChar> ContainsToken<C> for [char; LEN] { |
1972 | #[inline ] |
1973 | fn contains_token(&self, token: C) -> bool { |
1974 | let token: char = token.as_char(); |
1975 | self.iter().any(|t: &char| *t == token) |
1976 | } |
1977 | } |
1978 | |
1979 | impl<T> ContainsToken<T> for () { |
1980 | #[inline (always)] |
1981 | fn contains_token(&self, _token: T) -> bool { |
1982 | false |
1983 | } |
1984 | } |
1985 | |
1986 | macro_rules! impl_contains_token_for_tuple { |
1987 | ($($haystack:ident),+) => ( |
1988 | #[allow(non_snake_case)] |
1989 | impl<T, $($haystack),+> ContainsToken<T> for ($($haystack),+,) |
1990 | where |
1991 | T: Clone, |
1992 | $($haystack: ContainsToken<T>),+ |
1993 | { |
1994 | #[inline] |
1995 | fn contains_token(&self, token: T) -> bool { |
1996 | let ($(ref $haystack),+,) = *self; |
1997 | $($haystack.contains_token(token.clone()) || )+ false |
1998 | } |
1999 | } |
2000 | ) |
2001 | } |
2002 | |
2003 | macro_rules! impl_contains_token_for_tuples { |
2004 | ($haystack1:ident, $($haystack:ident),+) => { |
2005 | impl_contains_token_for_tuples!(__impl $haystack1; $($haystack),+); |
2006 | }; |
2007 | (__impl $($haystack:ident),+; $haystack1:ident $(,$haystack2:ident)*) => { |
2008 | impl_contains_token_for_tuple!($($haystack),+); |
2009 | impl_contains_token_for_tuples!(__impl $($haystack),+, $haystack1; $($haystack2),*); |
2010 | }; |
2011 | (__impl $($haystack:ident),+;) => { |
2012 | impl_contains_token_for_tuple!($($haystack),+); |
2013 | } |
2014 | } |
2015 | |
2016 | impl_contains_token_for_tuples!( |
2017 | F1, F2, F3, F4, F5, F6, F7, F8, F9, F10, F11, F12, F13, F14, F15, F16, F17, F18, F19, F20, F21 |
2018 | ); |
2019 | |
2020 | #[cfg (feature = "simd" )] |
2021 | #[inline (always)] |
2022 | fn memchr(token: u8, slice: &[u8]) -> Option<usize> { |
2023 | memchr::memchr(token, slice) |
2024 | } |
2025 | |
2026 | #[cfg (feature = "simd" )] |
2027 | #[inline (always)] |
2028 | fn memchr2(token: (u8, u8), slice: &[u8]) -> Option<usize> { |
2029 | memchr::memchr2(token.0, token.1, slice) |
2030 | } |
2031 | |
2032 | #[cfg (feature = "simd" )] |
2033 | #[inline (always)] |
2034 | fn memchr3(token: (u8, u8, u8), slice: &[u8]) -> Option<usize> { |
2035 | memchr::memchr3(token.0, token.1, token.2, slice) |
2036 | } |
2037 | |
2038 | #[cfg (not(feature = "simd" ))] |
2039 | #[inline (always)] |
2040 | fn memchr(token: u8, slice: &[u8]) -> Option<usize> { |
2041 | slice.iter().position(|t: &u8| *t == token) |
2042 | } |
2043 | |
2044 | #[cfg (not(feature = "simd" ))] |
2045 | #[inline (always)] |
2046 | fn memchr2(token: (u8, u8), slice: &[u8]) -> Option<usize> { |
2047 | slice.iter().position(|t: &u8| *t == token.0 || *t == token.1) |
2048 | } |
2049 | |
2050 | #[cfg (not(feature = "simd" ))] |
2051 | #[inline (always)] |
2052 | fn memchr3(token: (u8, u8, u8), slice: &[u8]) -> Option<usize> { |
2053 | sliceIter<'_, u8> |
2054 | .iter() |
2055 | .position(|t: &u8| *t == token.0 || *t == token.1 || *t == token.2) |
2056 | } |
2057 | |
2058 | #[inline (always)] |
2059 | fn memmem(slice: &[u8], literal: &[u8]) -> Option<crate::lib::std::ops::Range<usize>> { |
2060 | match literal.len() { |
2061 | 0 => Some(0..0), |
2062 | 1 => memchr(token:literal[0], slice).map(|i: usize| i..i + 1), |
2063 | _ => memmem_(slice, literal), |
2064 | } |
2065 | } |
2066 | |
2067 | #[inline (always)] |
2068 | fn memmem2(slice: &[u8], literal: (&[u8], &[u8])) -> Option<crate::lib::std::ops::Range<usize>> { |
2069 | match (literal.0.len(), literal.1.len()) { |
2070 | (0, _) | (_, 0) => Some(0..0), |
2071 | (1, 1) => memchr2((literal.0[0], literal.1[0]), slice).map(|i: usize| i..i + 1), |
2072 | _ => memmem2_(slice, literal), |
2073 | } |
2074 | } |
2075 | |
2076 | #[inline (always)] |
2077 | fn memmem3( |
2078 | slice: &[u8], |
2079 | literal: (&[u8], &[u8], &[u8]), |
2080 | ) -> Option<crate::lib::std::ops::Range<usize>> { |
2081 | match (literal.0.len(), literal.1.len(), literal.2.len()) { |
2082 | (0, _, _) | (_, 0, _) | (_, _, 0) => Some(0..0), |
2083 | (1, 1, 1) => memchr3((literal.0[0], literal.1[0], literal.2[0]), slice).map(|i: usize| i..i + 1), |
2084 | _ => memmem3_(slice, literal), |
2085 | } |
2086 | } |
2087 | |
2088 | #[cfg (feature = "simd" )] |
2089 | #[inline (always)] |
2090 | fn memmem_(slice: &[u8], literal: &[u8]) -> Option<crate::lib::std::ops::Range<usize>> { |
2091 | let &prefix = match literal.first() { |
2092 | Some(x) => x, |
2093 | None => return Some(0..0), |
2094 | }; |
2095 | #[allow (clippy::manual_find)] // faster this way |
2096 | for i in memchr::memchr_iter(prefix, slice) { |
2097 | if slice[i..].starts_with(literal) { |
2098 | let i_end = i + literal.len(); |
2099 | return Some(i..i_end); |
2100 | } |
2101 | } |
2102 | None |
2103 | } |
2104 | |
2105 | #[cfg (feature = "simd" )] |
2106 | fn memmem2_(slice: &[u8], literal: (&[u8], &[u8])) -> Option<crate::lib::std::ops::Range<usize>> { |
2107 | let prefix = match (literal.0.first(), literal.1.first()) { |
2108 | (Some(&a), Some(&b)) => (a, b), |
2109 | _ => return Some(0..0), |
2110 | }; |
2111 | #[allow (clippy::manual_find)] // faster this way |
2112 | for i in memchr::memchr2_iter(prefix.0, prefix.1, slice) { |
2113 | let subslice = &slice[i..]; |
2114 | if subslice.starts_with(literal.0) { |
2115 | let i_end = i + literal.0.len(); |
2116 | return Some(i..i_end); |
2117 | } |
2118 | if subslice.starts_with(literal.1) { |
2119 | let i_end = i + literal.1.len(); |
2120 | return Some(i..i_end); |
2121 | } |
2122 | } |
2123 | None |
2124 | } |
2125 | |
2126 | #[cfg (feature = "simd" )] |
2127 | fn memmem3_( |
2128 | slice: &[u8], |
2129 | literal: (&[u8], &[u8], &[u8]), |
2130 | ) -> Option<crate::lib::std::ops::Range<usize>> { |
2131 | let prefix = match (literal.0.first(), literal.1.first(), literal.2.first()) { |
2132 | (Some(&a), Some(&b), Some(&c)) => (a, b, c), |
2133 | _ => return Some(0..0), |
2134 | }; |
2135 | #[allow (clippy::manual_find)] // faster this way |
2136 | for i in memchr::memchr3_iter(prefix.0, prefix.1, prefix.2, slice) { |
2137 | let subslice = &slice[i..]; |
2138 | if subslice.starts_with(literal.0) { |
2139 | let i_end = i + literal.0.len(); |
2140 | return Some(i..i_end); |
2141 | } |
2142 | if subslice.starts_with(literal.1) { |
2143 | let i_end = i + literal.1.len(); |
2144 | return Some(i..i_end); |
2145 | } |
2146 | if subslice.starts_with(literal.2) { |
2147 | let i_end = i + literal.2.len(); |
2148 | return Some(i..i_end); |
2149 | } |
2150 | } |
2151 | None |
2152 | } |
2153 | |
2154 | #[cfg (not(feature = "simd" ))] |
2155 | fn memmem_(slice: &[u8], literal: &[u8]) -> Option<crate::lib::std::ops::Range<usize>> { |
2156 | for i: usize in 0..slice.len() { |
2157 | let subslice: &[u8] = &slice[i..]; |
2158 | if subslice.starts_with(needle:literal) { |
2159 | let i_end: usize = i + literal.len(); |
2160 | return Some(i..i_end); |
2161 | } |
2162 | } |
2163 | None |
2164 | } |
2165 | |
2166 | #[cfg (not(feature = "simd" ))] |
2167 | fn memmem2_(slice: &[u8], literal: (&[u8], &[u8])) -> Option<crate::lib::std::ops::Range<usize>> { |
2168 | for i: usize in 0..slice.len() { |
2169 | let subslice: &[u8] = &slice[i..]; |
2170 | if subslice.starts_with(needle:literal.0) { |
2171 | let i_end: usize = i + literal.0.len(); |
2172 | return Some(i..i_end); |
2173 | } |
2174 | if subslice.starts_with(needle:literal.1) { |
2175 | let i_end: usize = i + literal.1.len(); |
2176 | return Some(i..i_end); |
2177 | } |
2178 | } |
2179 | None |
2180 | } |
2181 | |
2182 | #[cfg (not(feature = "simd" ))] |
2183 | fn memmem3_( |
2184 | slice: &[u8], |
2185 | literal: (&[u8], &[u8], &[u8]), |
2186 | ) -> Option<crate::lib::std::ops::Range<usize>> { |
2187 | for i: usize in 0..slice.len() { |
2188 | let subslice: &[u8] = &slice[i..]; |
2189 | if subslice.starts_with(needle:literal.0) { |
2190 | let i_end: usize = i + literal.0.len(); |
2191 | return Some(i..i_end); |
2192 | } |
2193 | if subslice.starts_with(needle:literal.1) { |
2194 | let i_end: usize = i + literal.1.len(); |
2195 | return Some(i..i_end); |
2196 | } |
2197 | if subslice.starts_with(needle:literal.2) { |
2198 | let i_end: usize = i + literal.2.len(); |
2199 | return Some(i..i_end); |
2200 | } |
2201 | } |
2202 | None |
2203 | } |
2204 | |