regex_syntax/utf8.rs
1/*!
2Converts ranges of Unicode scalar values to equivalent ranges of UTF-8 bytes.
3
4This is sub-module is useful for constructing byte based automatons that need
5to embed UTF-8 decoding. The most common use of this module is in conjunction
6with the [`hir::ClassUnicodeRange`](../hir/struct.ClassUnicodeRange.html) type.
7
8See the documentation on the `Utf8Sequences` iterator for more details and
9an example.
10
11# Wait, what is this?
12
13This is simplest to explain with an example. Let's say you wanted to test
14whether a particular byte sequence was a Cyrillic character. One possible
15scalar value range is `[0400-04FF]`. The set of allowed bytes for this
16range can be expressed as a sequence of byte ranges:
17
18```text
19[D0-D3][80-BF]
20```
21
22This is simple enough: simply encode the boundaries, `0400` encodes to
23`D0 80` and `04FF` encodes to `D3 BF`, and create ranges from each
24corresponding pair of bytes: `D0` to `D3` and `80` to `BF`.
25
26However, what if you wanted to add the Cyrillic Supplementary characters to
27your range? Your range might then become `[0400-052F]`. The same procedure
28as above doesn't quite work because `052F` encodes to `D4 AF`. The byte ranges
29you'd get from the previous transformation would be `[D0-D4][80-AF]`. However,
30this isn't quite correct because this range doesn't capture many characters,
31for example, `04FF` (because its last byte, `BF` isn't in the range `80-AF`).
32
33Instead, you need multiple sequences of byte ranges:
34
35```text
36[D0-D3][80-BF] # matches codepoints 0400-04FF
37[D4][80-AF] # matches codepoints 0500-052F
38```
39
40This gets even more complicated if you want bigger ranges, particularly if
41they naively contain surrogate codepoints. For example, the sequence of byte
42ranges for the basic multilingual plane (`[0000-FFFF]`) look like this:
43
44```text
45[0-7F]
46[C2-DF][80-BF]
47[E0][A0-BF][80-BF]
48[E1-EC][80-BF][80-BF]
49[ED][80-9F][80-BF]
50[EE-EF][80-BF][80-BF]
51```
52
53Note that the byte ranges above will *not* match any erroneous encoding of
54UTF-8, including encodings of surrogate codepoints.
55
56And, of course, for all of Unicode (`[000000-10FFFF]`):
57
58```text
59[0-7F]
60[C2-DF][80-BF]
61[E0][A0-BF][80-BF]
62[E1-EC][80-BF][80-BF]
63[ED][80-9F][80-BF]
64[EE-EF][80-BF][80-BF]
65[F0][90-BF][80-BF][80-BF]
66[F1-F3][80-BF][80-BF][80-BF]
67[F4][80-8F][80-BF][80-BF]
68```
69
70This module automates the process of creating these byte ranges from ranges of
71Unicode scalar values.
72
73# Lineage
74
75I got the idea and general implementation strategy from Russ Cox in his
76[article on regexps](https://web.archive.org/web/20160404141123/https://swtch.com/~rsc/regexp/regexp3.html) and RE2.
77Russ Cox got it from Ken Thompson's `grep` (no source, folk lore?).
78I also got the idea from
79[Lucene](https://github.com/apache/lucene-solr/blob/ae93f4e7ac6a3908046391de35d4f50a0d3c59ca/lucene/core/src/java/org/apache/lucene/util/automaton/UTF32ToUTF8.java),
80which uses it for executing automata on their term index.
81*/
82
83#![deny(missing_docs)]
84
85use std::char;
86use std::fmt;
87use std::iter::FusedIterator;
88use std::slice;
89
90const MAX_UTF8_BYTES: usize = 4;
91
92/// Utf8Sequence represents a sequence of byte ranges.
93///
94/// To match a Utf8Sequence, a candidate byte sequence must match each
95/// successive range.
96///
97/// For example, if there are two ranges, `[C2-DF][80-BF]`, then the byte
98/// sequence `\xDD\x61` would not match because `0x61 < 0x80`.
99#[derive(Copy, Clone, Eq, PartialEq, PartialOrd, Ord)]
100pub enum Utf8Sequence {
101 /// One byte range.
102 One(Utf8Range),
103 /// Two successive byte ranges.
104 Two([Utf8Range; 2]),
105 /// Three successive byte ranges.
106 Three([Utf8Range; 3]),
107 /// Four successive byte ranges.
108 Four([Utf8Range; 4]),
109}
110
111impl Utf8Sequence {
112 /// Creates a new UTF-8 sequence from the encoded bytes of a scalar value
113 /// range.
114 ///
115 /// This assumes that `start` and `end` have the same length.
116 fn from_encoded_range(start: &[u8], end: &[u8]) -> Self {
117 assert_eq!(start.len(), end.len());
118 match start.len() {
119 2 => Utf8Sequence::Two([
120 Utf8Range::new(start[0], end[0]),
121 Utf8Range::new(start[1], end[1]),
122 ]),
123 3 => Utf8Sequence::Three([
124 Utf8Range::new(start[0], end[0]),
125 Utf8Range::new(start[1], end[1]),
126 Utf8Range::new(start[2], end[2]),
127 ]),
128 4 => Utf8Sequence::Four([
129 Utf8Range::new(start[0], end[0]),
130 Utf8Range::new(start[1], end[1]),
131 Utf8Range::new(start[2], end[2]),
132 Utf8Range::new(start[3], end[3]),
133 ]),
134 n => unreachable!("invalid encoded length: {}", n),
135 }
136 }
137
138 /// Returns the underlying sequence of byte ranges as a slice.
139 pub fn as_slice(&self) -> &[Utf8Range] {
140 use self::Utf8Sequence::*;
141 match *self {
142 One(ref r) => slice::from_ref(r),
143 Two(ref r) => &r[..],
144 Three(ref r) => &r[..],
145 Four(ref r) => &r[..],
146 }
147 }
148
149 /// Returns the number of byte ranges in this sequence.
150 ///
151 /// The length is guaranteed to be in the closed interval `[1, 4]`.
152 pub fn len(&self) -> usize {
153 self.as_slice().len()
154 }
155
156 /// Reverses the ranges in this sequence.
157 ///
158 /// For example, if this corresponds to the following sequence:
159 ///
160 /// ```text
161 /// [D0-D3][80-BF]
162 /// ```
163 ///
164 /// Then after reversal, it will be
165 ///
166 /// ```text
167 /// [80-BF][D0-D3]
168 /// ```
169 ///
170 /// This is useful when one is constructing a UTF-8 automaton to match
171 /// character classes in reverse.
172 pub fn reverse(&mut self) {
173 match *self {
174 Utf8Sequence::One(_) => {}
175 Utf8Sequence::Two(ref mut x) => x.reverse(),
176 Utf8Sequence::Three(ref mut x) => x.reverse(),
177 Utf8Sequence::Four(ref mut x) => x.reverse(),
178 }
179 }
180
181 /// Returns true if and only if a prefix of `bytes` matches this sequence
182 /// of byte ranges.
183 pub fn matches(&self, bytes: &[u8]) -> bool {
184 if bytes.len() < self.len() {
185 return false;
186 }
187 for (&b, r) in bytes.iter().zip(self) {
188 if !r.matches(b) {
189 return false;
190 }
191 }
192 true
193 }
194}
195
196impl<'a> IntoIterator for &'a Utf8Sequence {
197 type IntoIter = slice::Iter<'a, Utf8Range>;
198 type Item = &'a Utf8Range;
199
200 fn into_iter(self) -> Self::IntoIter {
201 self.as_slice().iter()
202 }
203}
204
205impl fmt::Debug for Utf8Sequence {
206 fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
207 use self::Utf8Sequence::*;
208 match *self {
209 One(ref r) => write!(f, "{:?}", r),
210 Two(ref r) => write!(f, "{:?}{:?}", r[0], r[1]),
211 Three(ref r) => write!(f, "{:?}{:?}{:?}", r[0], r[1], r[2]),
212 Four(ref r) => {
213 write!(f, "{:?}{:?}{:?}{:?}", r[0], r[1], r[2], r[3])
214 }
215 }
216 }
217}
218
219/// A single inclusive range of UTF-8 bytes.
220#[derive(Clone, Copy, Eq, PartialEq, PartialOrd, Ord)]
221pub struct Utf8Range {
222 /// Start of byte range (inclusive).
223 pub start: u8,
224 /// End of byte range (inclusive).
225 pub end: u8,
226}
227
228impl Utf8Range {
229 fn new(start: u8, end: u8) -> Self {
230 Utf8Range { start, end }
231 }
232
233 /// Returns true if and only if the given byte is in this range.
234 pub fn matches(&self, b: u8) -> bool {
235 self.start <= b && b <= self.end
236 }
237}
238
239impl fmt::Debug for Utf8Range {
240 fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
241 if self.start == self.end {
242 write!(f, "[{:X}]", self.start)
243 } else {
244 write!(f, "[{:X}-{:X}]", self.start, self.end)
245 }
246 }
247}
248
249/// An iterator over ranges of matching UTF-8 byte sequences.
250///
251/// The iteration represents an alternation of comprehensive byte sequences
252/// that match precisely the set of UTF-8 encoded scalar values.
253///
254/// A byte sequence corresponds to one of the scalar values in the range given
255/// if and only if it completely matches exactly one of the sequences of byte
256/// ranges produced by this iterator.
257///
258/// Each sequence of byte ranges matches a unique set of bytes. That is, no two
259/// sequences will match the same bytes.
260///
261/// # Example
262///
263/// This shows how to match an arbitrary byte sequence against a range of
264/// scalar values.
265///
266/// ```rust
267/// use regex_syntax::utf8::{Utf8Sequences, Utf8Sequence};
268///
269/// fn matches(seqs: &[Utf8Sequence], bytes: &[u8]) -> bool {
270/// for range in seqs {
271/// if range.matches(bytes) {
272/// return true;
273/// }
274/// }
275/// false
276/// }
277///
278/// // Test the basic multilingual plane.
279/// let seqs: Vec<_> = Utf8Sequences::new('\u{0}', '\u{FFFF}').collect();
280///
281/// // UTF-8 encoding of 'a'.
282/// assert!(matches(&seqs, &[0x61]));
283/// // UTF-8 encoding of '☃' (`\u{2603}`).
284/// assert!(matches(&seqs, &[0xE2, 0x98, 0x83]));
285/// // UTF-8 encoding of `\u{10348}` (outside the BMP).
286/// assert!(!matches(&seqs, &[0xF0, 0x90, 0x8D, 0x88]));
287/// // Tries to match against a UTF-8 encoding of a surrogate codepoint,
288/// // which is invalid UTF-8, and therefore fails, despite the fact that
289/// // the corresponding codepoint (0xD800) falls in the range given.
290/// assert!(!matches(&seqs, &[0xED, 0xA0, 0x80]));
291/// // And fails against plain old invalid UTF-8.
292/// assert!(!matches(&seqs, &[0xFF, 0xFF]));
293/// ```
294///
295/// If this example seems circuitous, that's because it is! It's meant to be
296/// illustrative. In practice, you could just try to decode your byte sequence
297/// and compare it with the scalar value range directly. However, this is not
298/// always possible (for example, in a byte based automaton).
299#[derive(Debug)]
300pub struct Utf8Sequences {
301 range_stack: Vec<ScalarRange>,
302}
303
304impl Utf8Sequences {
305 /// Create a new iterator over UTF-8 byte ranges for the scalar value range
306 /// given.
307 pub fn new(start: char, end: char) -> Self {
308 let mut it = Utf8Sequences { range_stack: vec![] };
309 it.push(start as u32, end as u32);
310 it
311 }
312
313 /// reset resets the scalar value range.
314 /// Any existing state is cleared, but resources may be reused.
315 ///
316 /// N.B. Benchmarks say that this method is dubious.
317 #[doc(hidden)]
318 pub fn reset(&mut self, start: char, end: char) {
319 self.range_stack.clear();
320 self.push(start as u32, end as u32);
321 }
322
323 fn push(&mut self, start: u32, end: u32) {
324 self.range_stack.push(ScalarRange { start, end });
325 }
326}
327
328struct ScalarRange {
329 start: u32,
330 end: u32,
331}
332
333impl fmt::Debug for ScalarRange {
334 fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
335 write!(f, "ScalarRange({:X}, {:X})", self.start, self.end)
336 }
337}
338
339impl Iterator for Utf8Sequences {
340 type Item = Utf8Sequence;
341
342 fn next(&mut self) -> Option<Self::Item> {
343 'TOP: while let Some(mut r) = self.range_stack.pop() {
344 'INNER: loop {
345 if let Some((r1, r2)) = r.split() {
346 self.push(r2.start, r2.end);
347 r.start = r1.start;
348 r.end = r1.end;
349 continue 'INNER;
350 }
351 if !r.is_valid() {
352 continue 'TOP;
353 }
354 for i in 1..MAX_UTF8_BYTES {
355 let max = max_scalar_value(i);
356 if r.start <= max && max < r.end {
357 self.push(max + 1, r.end);
358 r.end = max;
359 continue 'INNER;
360 }
361 }
362 if let Some(ascii_range) = r.as_ascii() {
363 return Some(Utf8Sequence::One(ascii_range));
364 }
365 for i in 1..MAX_UTF8_BYTES {
366 let m = (1 << (6 * i)) - 1;
367 if (r.start & !m) != (r.end & !m) {
368 if (r.start & m) != 0 {
369 self.push((r.start | m) + 1, r.end);
370 r.end = r.start | m;
371 continue 'INNER;
372 }
373 if (r.end & m) != m {
374 self.push(r.end & !m, r.end);
375 r.end = (r.end & !m) - 1;
376 continue 'INNER;
377 }
378 }
379 }
380 let mut start = [0; MAX_UTF8_BYTES];
381 let mut end = [0; MAX_UTF8_BYTES];
382 let n = r.encode(&mut start, &mut end);
383 return Some(Utf8Sequence::from_encoded_range(
384 &start[0..n],
385 &end[0..n],
386 ));
387 }
388 }
389 None
390 }
391}
392
393impl FusedIterator for Utf8Sequences {}
394
395impl ScalarRange {
396 /// split splits this range if it overlaps with a surrogate codepoint.
397 ///
398 /// Either or both ranges may be invalid.
399 fn split(&self) -> Option<(ScalarRange, ScalarRange)> {
400 if self.start < 0xE000 && self.end > 0xD7FF {
401 Some((
402 ScalarRange { start: self.start, end: 0xD7FF },
403 ScalarRange { start: 0xE000, end: self.end },
404 ))
405 } else {
406 None
407 }
408 }
409
410 /// is_valid returns true if and only if start <= end.
411 fn is_valid(&self) -> bool {
412 self.start <= self.end
413 }
414
415 /// as_ascii returns this range as a Utf8Range if and only if all scalar
416 /// values in this range can be encoded as a single byte.
417 fn as_ascii(&self) -> Option<Utf8Range> {
418 if self.is_ascii() {
419 Some(Utf8Range::new(self.start as u8, self.end as u8))
420 } else {
421 None
422 }
423 }
424
425 /// is_ascii returns true if the range is ASCII only (i.e., takes a single
426 /// byte to encode any scalar value).
427 fn is_ascii(&self) -> bool {
428 self.is_valid() && self.end <= 0x7f
429 }
430
431 /// encode writes the UTF-8 encoding of the start and end of this range
432 /// to the corresponding destination slices, and returns the number of
433 /// bytes written.
434 ///
435 /// The slices should have room for at least `MAX_UTF8_BYTES`.
436 fn encode(&self, start: &mut [u8], end: &mut [u8]) -> usize {
437 let cs = char::from_u32(self.start).unwrap();
438 let ce = char::from_u32(self.end).unwrap();
439 let ss = cs.encode_utf8(start);
440 let se = ce.encode_utf8(end);
441 assert_eq!(ss.len(), se.len());
442 ss.len()
443 }
444}
445
446fn max_scalar_value(nbytes: usize) -> u32 {
447 match nbytes {
448 1 => 0x007F,
449 2 => 0x07FF,
450 3 => 0xFFFF,
451 4 => 0x0010_FFFF,
452 _ => unreachable!("invalid UTF-8 byte sequence size"),
453 }
454}
455
456#[cfg(test)]
457mod tests {
458 use std::char;
459
460 use crate::utf8::{Utf8Range, Utf8Sequences};
461
462 fn rutf8(s: u8, e: u8) -> Utf8Range {
463 Utf8Range::new(s, e)
464 }
465
466 fn never_accepts_surrogate_codepoints(start: char, end: char) {
467 for cp in 0xD800..0xE000 {
468 let buf = encode_surrogate(cp);
469 for r in Utf8Sequences::new(start, end) {
470 if r.matches(&buf) {
471 panic!(
472 "Sequence ({:X}, {:X}) contains range {:?}, \
473 which matches surrogate code point {:X} \
474 with encoded bytes {:?}",
475 start as u32, end as u32, r, cp, buf,
476 );
477 }
478 }
479 }
480 }
481
482 #[test]
483 fn codepoints_no_surrogates() {
484 never_accepts_surrogate_codepoints('\u{0}', '\u{FFFF}');
485 never_accepts_surrogate_codepoints('\u{0}', '\u{10FFFF}');
486 never_accepts_surrogate_codepoints('\u{0}', '\u{10FFFE}');
487 never_accepts_surrogate_codepoints('\u{80}', '\u{10FFFF}');
488 never_accepts_surrogate_codepoints('\u{D7FF}', '\u{E000}');
489 }
490
491 #[test]
492 fn single_codepoint_one_sequence() {
493 // Tests that every range of scalar values that contains a single
494 // scalar value is recognized by one sequence of byte ranges.
495 for i in 0x0..=0x0010_FFFF {
496 let c = match char::from_u32(i) {
497 None => continue,
498 Some(c) => c,
499 };
500 let seqs: Vec<_> = Utf8Sequences::new(c, c).collect();
501 assert_eq!(seqs.len(), 1);
502 }
503 }
504
505 #[test]
506 fn bmp() {
507 use crate::utf8::Utf8Sequence::*;
508
509 let seqs = Utf8Sequences::new('\u{0}', '\u{FFFF}').collect::<Vec<_>>();
510 assert_eq!(
511 seqs,
512 vec![
513 One(rutf8(0x0, 0x7F)),
514 Two([rutf8(0xC2, 0xDF), rutf8(0x80, 0xBF)]),
515 Three([
516 rutf8(0xE0, 0xE0),
517 rutf8(0xA0, 0xBF),
518 rutf8(0x80, 0xBF)
519 ]),
520 Three([
521 rutf8(0xE1, 0xEC),
522 rutf8(0x80, 0xBF),
523 rutf8(0x80, 0xBF)
524 ]),
525 Three([
526 rutf8(0xED, 0xED),
527 rutf8(0x80, 0x9F),
528 rutf8(0x80, 0xBF)
529 ]),
530 Three([
531 rutf8(0xEE, 0xEF),
532 rutf8(0x80, 0xBF),
533 rutf8(0x80, 0xBF)
534 ]),
535 ]
536 );
537 }
538
539 #[test]
540 fn reverse() {
541 use crate::utf8::Utf8Sequence::*;
542
543 let mut s = One(rutf8(0xA, 0xB));
544 s.reverse();
545 assert_eq!(s.as_slice(), &[rutf8(0xA, 0xB)]);
546
547 let mut s = Two([rutf8(0xA, 0xB), rutf8(0xB, 0xC)]);
548 s.reverse();
549 assert_eq!(s.as_slice(), &[rutf8(0xB, 0xC), rutf8(0xA, 0xB)]);
550
551 let mut s = Three([rutf8(0xA, 0xB), rutf8(0xB, 0xC), rutf8(0xC, 0xD)]);
552 s.reverse();
553 assert_eq!(
554 s.as_slice(),
555 &[rutf8(0xC, 0xD), rutf8(0xB, 0xC), rutf8(0xA, 0xB)]
556 );
557
558 let mut s = Four([
559 rutf8(0xA, 0xB),
560 rutf8(0xB, 0xC),
561 rutf8(0xC, 0xD),
562 rutf8(0xD, 0xE),
563 ]);
564 s.reverse();
565 assert_eq!(
566 s.as_slice(),
567 &[
568 rutf8(0xD, 0xE),
569 rutf8(0xC, 0xD),
570 rutf8(0xB, 0xC),
571 rutf8(0xA, 0xB)
572 ]
573 );
574 }
575
576 fn encode_surrogate(cp: u32) -> [u8; 3] {
577 const TAG_CONT: u8 = 0b1000_0000;
578 const TAG_THREE_B: u8 = 0b1110_0000;
579
580 assert!(0xD800 <= cp && cp < 0xE000);
581 let mut dst = [0; 3];
582 dst[0] = (cp >> 12 & 0x0F) as u8 | TAG_THREE_B;
583 dst[1] = (cp >> 6 & 0x3F) as u8 | TAG_CONT;
584 dst[2] = (cp & 0x3F) as u8 | TAG_CONT;
585 dst
586 }
587}