1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
//! ## Diff library with semantic cleanup, based on Google's diff-match-patch
//!
//! This library is a port of the Diff component of [Diff Match Patch] to Rust.
//! The diff implementation is based on [Myers' diff algorithm] but includes
//! some [semantic cleanups] to increase human readability by factoring out
//! commonalities which are likely to be coincidental.
//!
//! Diff Match Patch was originally built in 2006 to power Google Docs.
//!
//! # Interface
//!
//! Here is the entire API of the Rust implementation. It operates on borrowed
//! strings and the return value of the diff algorithm is a vector of chunks
//! pointing into slices of those input strings.
//!
//! ```
//! pub enum Chunk<'a> {
//!     Equal(&'a str),
//!     Delete(&'a str),
//!     Insert(&'a str),
//! }
//!
//! # const IGNORE: &str = stringify! {
//! pub fn diff(text1: &str, text2: &str) -> Vec<Chunk>;
//! # };
//! ```
//!
//! [Diff Match Patch]: https://github.com/google/diff-match-patch
//! [Myers' diff algorithm]: https://neil.fraser.name/writing/diff/myers.pdf
//! [semantic cleanups]: https://neil.fraser.name/writing/diff/

#![doc(html_root_url = "https://docs.rs/dissimilar/1.0.1")]
#![allow(
    clippy::block_in_if_condition_stmt,
    clippy::collapsible_if,
    clippy::comparison_chain,
    clippy::new_without_default
)]

mod find;
mod range;

#[cfg(test)]
mod tests;

use crate::range::{bytes, str, Range};
use std::cmp;
use std::collections::VecDeque;
use std::fmt::{self, Debug};

#[derive(Copy, Clone, PartialEq, Eq)]
pub enum Chunk<'a> {
    Equal(&'a str),
    Delete(&'a str),
    Insert(&'a str),
}

#[derive(Copy, Clone)]
enum Diff<'a, 'b> {
    Equal(Range<'a>, Range<'b>),
    Delete(Range<'a>),
    Insert(Range<'b>),
}

impl<'tmp, 'a: 'tmp, 'b: 'tmp> Diff<'a, 'b> {
    fn text(&self) -> Range<'tmp> {
        match *self {
            Diff::Equal(range, _) | Diff::Delete(range) | Diff::Insert(range) => range,
        }
    }

    fn grow_left(&mut self, increment: usize) {
        self.for_each(|range| {
            range.offset -= increment;
            range.len += increment;
        });
    }

    fn grow_right(&mut self, increment: usize) {
        self.for_each(|range| range.len += increment);
    }

    fn shift_left(&mut self, increment: usize) {
        self.for_each(|range| range.offset -= increment);
    }

    fn shift_right(&mut self, increment: usize) {
        self.for_each(|range| range.offset += increment);
    }

    fn for_each(&mut self, f: impl Fn(&mut Range)) {
        match self {
            Diff::Equal(range1, range2) => {
                f(range1);
                f(range2);
            }
            Diff::Delete(range) => f(range),
            Diff::Insert(range) => f(range),
        }
    }
}

pub fn diff<'a>(text1: &'a str, text2: &'a str) -> Vec<Chunk<'a>> {
    let text1 = Range::new(text1, ..);
    let text2 = Range::new(text2, ..);
    let mut solution = main(text1, text2);
    cleanup_char_boundary(&mut solution);
    cleanup_semantic(&mut solution);
    cleanup_merge(&mut solution);
    solution.diffs.into_iter().map(Chunk::from).collect()
}

struct Solution<'a, 'b> {
    text1: Range<'a>,
    text2: Range<'b>,
    diffs: Vec<Diff<'a, 'b>>,
    utf8: bool,
}

fn main<'a, 'b>(mut text1: Range<'a>, mut text2: Range<'b>) -> Solution<'a, 'b> {
    let whole1 = text1;
    let whole2 = text2;

    // Trim off common prefix.
    let common_prefix_len = common_prefix_bytes(text1, text2);
    let common_prefix = Diff::Equal(
        text1.substring(..common_prefix_len),
        text2.substring(..common_prefix_len),
    );
    text1 = text1.substring(common_prefix_len..);
    text2 = text2.substring(common_prefix_len..);

    // Trim off common suffix.
    let common_suffix_len = common_suffix_bytes(text1, text2);
    let common_suffix = Diff::Equal(
        text1.substring(text1.len - common_suffix_len..),
        text2.substring(text2.len - common_suffix_len..),
    );
    text1 = text1.substring(..text1.len - common_suffix_len);
    text2 = text2.substring(..text2.len - common_suffix_len);

    // Compute the diff on the middle block.
    let mut solution = Solution {
        text1: whole1,
        text2: whole2,
        diffs: compute(text1, text2),
        utf8: false,
    };

    // Restore the prefix and suffix.
    if common_prefix_len > 0 {
        solution.diffs.insert(0, common_prefix);
    }
    if common_suffix_len > 0 {
        solution.diffs.push(common_suffix);
    }

    cleanup_merge(&mut solution);

    solution
}

// Find the differences between two texts. Assumes that the texts do not have
// any common prefix or suffix.
fn compute<'a, 'b>(text1: Range<'a>, text2: Range<'b>) -> Vec<Diff<'a, 'b>> {
    match (text1.is_empty(), text2.is_empty()) {
        (true, true) => return Vec::new(),
        (true, false) => return vec![Diff::Insert(text2)],
        (false, true) => return vec![Diff::Delete(text1)],
        (false, false) => {}
    }

    // Check for entire shorter text inside the longer text.
    if text1.len > text2.len {
        if let Some(i) = text1.find(text2) {
            return vec![
                Diff::Delete(text1.substring(..i)),
                Diff::Equal(text1.substring(i..i + text2.len), text2),
                Diff::Delete(text1.substring(i + text2.len..)),
            ];
        }
    } else {
        if let Some(i) = text2.find(text1) {
            return vec![
                Diff::Insert(text2.substring(..i)),
                Diff::Equal(text1, text2.substring(i..i + text1.len)),
                Diff::Insert(text2.substring(i + text1.len..)),
            ];
        }
    }

    if text1.len == 1 || text2.len == 1 {
        // Single character string.
        // After the previous check, the character can't be an equality.
        return vec![Diff::Delete(text1), Diff::Insert(text2)];
    }

    bisect(text1, text2)
}

// Find the 'middle snake' of a diff, split the problem in two and return the
// recursively constructed diff.
//
// See Myers 1986 paper: An O(ND) Difference Algorithm and Its Variations.
fn bisect<'a, 'b>(text1: Range<'a>, text2: Range<'b>) -> Vec<Diff<'a, 'b>> {
    let max_d = (text1.len + text2.len + 1) / 2;
    let v_offset = max_d;
    let v_len = 2 * max_d;
    let mut v1 = vec![-1isize; v_len];
    let mut v2 = vec![-1isize; v_len];
    v1[v_offset + 1] = 0;
    v2[v_offset + 1] = 0;
    let delta = text1.len as isize - text2.len as isize;
    // If the total number of characters is odd, then the front path will
    // collide with the reverse path.
    let front = delta % 2 != 0;
    // Offsets for start and end of k loop.
    // Prevents mapping of space beyond the grid.
    let mut k1start = 0;
    let mut k1end = 0;
    let mut k2start = 0;
    let mut k2end = 0;
    for d in 0..max_d as isize {
        // Walk the front path one step.
        let mut k1 = -d + k1start;
        while k1 <= d - k1end {
            let k1_offset = (v_offset as isize + k1) as usize;
            let mut x1 = if k1 == -d || (k1 != d && v1[k1_offset - 1] < v1[k1_offset + 1]) {
                v1[k1_offset + 1]
            } else {
                v1[k1_offset - 1] + 1
            } as usize;
            let mut y1 = (x1 as isize - k1) as usize;
            if let (Some(s1), Some(s2)) = (text1.get(x1..), text2.get(y1..)) {
                let advance = common_prefix_bytes(s1, s2);
                x1 += advance;
                y1 += advance;
            }
            v1[k1_offset] = x1 as isize;
            if x1 > text1.len {
                // Ran off the right of the graph.
                k1end += 2;
            } else if y1 > text2.len {
                // Ran off the bottom of the graph.
                k1start += 2;
            } else if front {
                let k2_offset = v_offset as isize + delta - k1;
                if k2_offset >= 0 && k2_offset < v_len as isize && v2[k2_offset as usize] != -1 {
                    // Mirror x2 onto top-left coordinate system.
                    let x2 = text1.len as isize - v2[k2_offset as usize];
                    if x1 as isize >= x2 {
                        // Overlap detected.
                        return bisect_split(text1, text2, x1, y1);
                    }
                }
            }
            k1 += 2;
        }

        // Walk the reverse path one step.
        let mut k2 = -d + k2start;
        while k2 <= d - k2end {
            let k2_offset = (v_offset as isize + k2) as usize;
            let mut x2 = if k2 == -d || (k2 != d && v2[k2_offset - 1] < v2[k2_offset + 1]) {
                v2[k2_offset + 1]
            } else {
                v2[k2_offset - 1] + 1
            } as usize;
            let mut y2 = (x2 as isize - k2) as usize;
            if x2 < text1.len && y2 < text2.len {
                let advance = common_suffix_bytes(
                    text1.substring(..text1.len - x2),
                    text2.substring(..text2.len - y2),
                );
                x2 += advance;
                y2 += advance;
            }
            v2[k2_offset] = x2 as isize;
            if x2 > text1.len {
                // Ran off the left of the graph.
                k2end += 2;
            } else if y2 > text2.len {
                // Ran off the top of the graph.
                k2start += 2;
            } else if !front {
                let k1_offset = v_offset as isize + delta - k2;
                if k1_offset >= 0 && k1_offset < v_len as isize && v1[k1_offset as usize] != -1 {
                    let x1 = v1[k1_offset as usize] as usize;
                    let y1 = v_offset + x1 - k1_offset as usize;
                    // Mirror x2 onto top-left coordinate system.
                    x2 = text1.len - x2;
                    if x1 >= x2 {
                        // Overlap detected.
                        return bisect_split(text1, text2, x1, y1);
                    }
                }
            }
            k2 += 2;
        }
    }
    // Number of diffs equals number of characters, no commonality at all.
    vec![Diff::Delete(text1), Diff::Insert(text2)]
}

// Given the location of the 'middle snake', split the diff in two parts and
// recurse.
fn bisect_split<'a, 'b>(
    text1: Range<'a>,
    text2: Range<'b>,
    x: usize,
    y: usize,
) -> Vec<Diff<'a, 'b>> {
    let (text1a, text1b) = text1.split_at(x);
    let (text2a, text2b) = text2.split_at(y);

    // Compute both diffs serially.
    let mut diffs = main(text1a, text2a).diffs;
    diffs.extend(main(text1b, text2b).diffs);

    diffs
}

// Determine the length of the common prefix of two strings.
fn common_prefix(text1: Range, text2: Range) -> usize {
    for ((i, ch1), ch2) in text1.char_indices().zip(text2.chars()) {
        if ch1 != ch2 {
            return i;
        }
    }
    cmp::min(text1.len, text2.len)
}

// Determine the length of the common suffix of two strings.
fn common_suffix(text1: Range, text2: Range) -> usize {
    for ((i, ch1), ch2) in text1.char_indices().rev().zip(text2.chars().rev()) {
        if ch1 != ch2 {
            return text1.len - i - ch1.len_utf8();
        }
    }
    cmp::min(text1.len, text2.len)
}

fn common_prefix_bytes(text1: Range, text2: Range) -> usize {
    for (i, (b1, b2)) in text1.bytes().zip(text2.bytes()).enumerate() {
        if b1 != b2 {
            return i;
        }
    }
    cmp::min(text1.len, text2.len)
}

fn common_suffix_bytes(text1: Range, text2: Range) -> usize {
    for (i, (b1, b2)) in text1.bytes().rev().zip(text2.bytes().rev()).enumerate() {
        if b1 != b2 {
            return i;
        }
    }
    cmp::min(text1.len, text2.len)
}

// Determine if the suffix of one string is the prefix of another.
//
// Returns the number of characters common to the end of the first string and
// the start of the second string.
fn common_overlap(mut text1: Range, mut text2: Range) -> usize {
    // Eliminate the null case.
    if text1.is_empty() || text2.is_empty() {
        return 0;
    }
    // Truncate the longer string.
    if text1.len > text2.len {
        text1 = text1.substring(text1.len - text2.len..);
    } else if text1.len < text2.len {
        text2 = text2.substring(..text1.len);
    }
    // Quick check for the worst case.
    if bytes(text1) == bytes(text2) {
        return text1.len;
    }

    // Start by looking for a single character match
    // and increase length until no match is found.
    // Performance analysis: https://neil.fraser.name/news/2010/11/04/
    let mut best = 0;
    let mut length = 1;
    loop {
        let pattern = text1.substring(text1.len - length..);
        let found = match text2.find(pattern) {
            Some(found) => found,
            None => return best,
        };
        length += found;
        if found == 0
            || bytes(text1.substring(text1.len - length..)) == bytes(text2.substring(..length))
        {
            best = length;
            length += 1;
        }
    }
}

fn cleanup_char_boundary(solution: &mut Solution) {
    fn boundary_down(doc: &str, pos: usize) -> usize {
        let mut adjust = 0;
        while !doc.is_char_boundary(pos - adjust) {
            adjust += 1;
        }
        adjust
    }

    fn boundary_up(doc: &str, pos: usize) -> usize {
        let mut adjust = 0;
        while !doc.is_char_boundary(pos + adjust) {
            adjust += 1;
        }
        adjust
    }

    for diff in &mut solution.diffs {
        match diff {
            Diff::Equal(range1, range2) => {
                let adjust = boundary_up(range1.doc, range1.offset);
                range1.offset += adjust;
                range1.len -= adjust;
                range2.offset += adjust;
                range2.len -= adjust;
                let adjust = boundary_down(range1.doc, range1.offset + range1.len);
                range1.len -= adjust;
                range2.len -= adjust;
            }
            Diff::Delete(range) => {
                let adjust = boundary_down(range.doc, range.offset);
                range.offset -= adjust;
                range.len += adjust;
                let adjust = boundary_up(range.doc, range.offset + range.len);
                range.len += adjust;
            }
            Diff::Insert(range) => {
                let adjust = boundary_down(range.doc, range.offset);
                range.offset -= adjust;
                range.len += adjust;
                let adjust = boundary_up(range.doc, range.offset + range.len);
                range.len += adjust;
            }
        }
    }

    solution.utf8 = true;
}

// Reduce the number of edits by eliminating semantically trivial equalities.
fn cleanup_semantic(solution: &mut Solution) {
    let mut diffs = &mut solution.diffs;
    if diffs.is_empty() {
        return;
    }

    let mut changes = false;
    let mut equalities = VecDeque::new(); // Double-ended queue of equalities.
    let mut last_equality = None; // Always equal to equalities.peek().text
    let mut pointer = 0;
    // Number of characters that changed prior to the equality.
    let mut len_insertions1 = 0;
    let mut len_deletions1 = 0;
    // Number of characters that changed after the equality.
    let mut len_insertions2 = 0;
    let mut len_deletions2 = 0;
    while let Some(&this_diff) = diffs.get(pointer) {
        match this_diff {
            Diff::Equal(text1, text2) => {
                equalities.push_back(pointer);
                len_insertions1 = len_insertions2;
                len_deletions1 = len_deletions2;
                len_insertions2 = 0;
                len_deletions2 = 0;
                last_equality = Some((text1, text2));
                pointer += 1;
                continue;
            }
            Diff::Delete(text) => len_deletions2 += text.len,
            Diff::Insert(text) => len_insertions2 += text.len,
        }
        // Eliminate an equality that is smaller or equal to the edits on both
        // sides of it.
        if last_equality.map_or(false, |(last_equality, _)| {
            last_equality.len <= cmp::max(len_insertions1, len_deletions1)
                && last_equality.len <= cmp::max(len_insertions2, len_deletions2)
        }) {
            // Jump back to offending equality.
            pointer = equalities.pop_back().unwrap();

            // Replace equality with a delete.
            diffs[pointer] = Diff::Delete(last_equality.unwrap().0);
            // Insert a corresponding insert.
            diffs.insert(pointer + 1, Diff::Insert(last_equality.unwrap().1));

            len_insertions1 = 0; // Reset the counters.
            len_insertions2 = 0;
            len_deletions1 = 0;
            len_deletions2 = 0;
            last_equality = None;
            changes = true;

            // Throw away the previous equality (it needs to be reevaluated).
            equalities.pop_back();
            if let Some(back) = equalities.back() {
                // There is a safe equality we can fall back to.
                pointer = *back;
            } else {
                // There are no previous equalities, jump back to the start.
                pointer = 0;
                continue;
            }
        }
        pointer += 1;
    }

    // Normalize the diff.
    if changes {
        cleanup_merge(solution);
    }
    cleanup_semantic_lossless(solution);
    diffs = &mut solution.diffs;

    // Find any overlaps between deletions and insertions.
    // e.g: <del>abcxxx</del><ins>xxxdef</ins>
    //   -> <del>abc</del>xxx<ins>def</ins>
    // e.g: <del>xxxabc</del><ins>defxxx</ins>
    //   -> <ins>def</ins>xxx<del>abc</del>
    // Only extract an overlap if it is as big as the edit ahead or behind it.
    let mut pointer = 1;
    while let Some(&this_diff) = diffs.get(pointer) {
        let prev_diff = diffs[pointer - 1];
        if let (Diff::Delete(deletion), Diff::Insert(insertion)) = (prev_diff, this_diff) {
            let overlap_len1 = common_overlap(deletion, insertion);
            let overlap_len2 = common_overlap(insertion, deletion);
            let overlap_min = cmp::min(deletion.len, insertion.len);
            if overlap_len1 >= overlap_len2 && 2 * overlap_len1 >= overlap_min {
                // Overlap found. Insert an equality and trim the surrounding edits.
                diffs.insert(
                    pointer,
                    Diff::Equal(
                        deletion.substring(deletion.len - overlap_len1..deletion.len),
                        insertion.substring(..overlap_len1),
                    ),
                );
                diffs[pointer - 1] =
                    Diff::Delete(deletion.substring(..deletion.len - overlap_len1));
                diffs[pointer + 1] = Diff::Insert(insertion.substring(overlap_len1..));
            } else if overlap_len1 < overlap_len2 && 2 * overlap_len2 >= overlap_min {
                // Reverse overlap found.
                // Insert an equality and swap and trim the surrounding edits.
                diffs.insert(
                    pointer,
                    Diff::Equal(
                        deletion.substring(..overlap_len2),
                        insertion.substring(insertion.len - overlap_len2..insertion.len),
                    ),
                );
                diffs[pointer - 1] =
                    Diff::Insert(insertion.substring(..insertion.len - overlap_len2));
                diffs[pointer + 1] = Diff::Delete(deletion.substring(overlap_len2..));
            }
            pointer += 1;
        }
        pointer += 1;
    }
}

// Look for single edits surrounded on both sides by equalities which can be
// shifted sideways to align the edit to a word boundary.
//
// e.g: The c<ins>at c</ins>ame. -> The <ins>cat </ins>came.
fn cleanup_semantic_lossless(solution: &mut Solution) {
    let diffs = &mut solution.diffs;
    let mut pointer = 1;
    while let Some(&next_diff) = diffs.get(pointer + 1) {
        let prev_diff = diffs[pointer - 1];
        if let (
            Diff::Equal(mut prev_equal1, mut prev_equal2),
            Diff::Equal(mut next_equal1, mut next_equal2),
        ) = (prev_diff, next_diff)
        {
            // This is a single edit surrounded by equalities.
            let mut edit = diffs[pointer];

            // First, shift the edit as far left as possible.
            let common_offset = common_suffix(prev_equal1, edit.text());
            let original_prev_len = prev_equal1.len;
            prev_equal1.len -= common_offset;
            prev_equal2.len -= common_offset;
            edit.shift_left(common_offset);
            next_equal1.offset -= common_offset;
            next_equal1.len += common_offset;
            next_equal2.offset -= common_offset;
            next_equal2.len += common_offset;

            // Second, step character by character right, looking for the best fit.
            let mut best_prev_equal = (prev_equal1, prev_equal2);
            let mut best_edit = edit;
            let mut best_next_equal = (next_equal1, next_equal2);
            let mut best_score = cleanup_semantic_score(prev_equal1, edit.text())
                + cleanup_semantic_score(edit.text(), next_equal1);
            while !edit.text().is_empty()
                && !next_equal1.is_empty()
                && edit.text().chars().next().unwrap() == next_equal1.chars().next().unwrap()
            {
                let increment = edit.text().chars().next().unwrap().len_utf8();
                prev_equal1.len += increment;
                prev_equal2.len += increment;
                edit.shift_right(increment);
                next_equal1.offset += increment;
                next_equal1.len -= increment;
                next_equal2.offset += increment;
                next_equal2.len -= increment;
                let score = cleanup_semantic_score(prev_equal1, edit.text())
                    + cleanup_semantic_score(edit.text(), next_equal1);
                // The >= encourages trailing rather than leading whitespace on edits.
                if score >= best_score {
                    best_score = score;
                    best_prev_equal = (prev_equal1, prev_equal2);
                    best_edit = edit;
                    best_next_equal = (next_equal1, next_equal2);
                }
            }

            if original_prev_len != best_prev_equal.0.len {
                // We have an improvement, save it back to the diff.
                if best_next_equal.0.is_empty() {
                    diffs.remove(pointer + 1);
                } else {
                    diffs[pointer + 1] = Diff::Equal(best_next_equal.0, best_next_equal.1);
                }
                diffs[pointer] = best_edit;
                if best_prev_equal.0.is_empty() {
                    diffs.remove(pointer - 1);
                    pointer -= 1;
                } else {
                    diffs[pointer - 1] = Diff::Equal(best_prev_equal.0, best_prev_equal.1);
                }
            }
        }
        pointer += 1;
    }
}

// Given two strings, compute a score representing whether the internal boundary
// falls on logical boundaries.
//
// Scores range from 6 (best) to 0 (worst).
fn cleanup_semantic_score(one: Range, two: Range) -> usize {
    if one.is_empty() || two.is_empty() {
        // Edges are the best.
        return 6;
    }

    // Each port of this function behaves slightly differently due to subtle
    // differences in each language's definition of things like 'whitespace'.
    // Since this function's purpose is largely cosmetic, the choice has been
    // made to use each language's native features rather than force total
    // conformity.
    let char1 = one.chars().next_back().unwrap();
    let char2 = two.chars().next().unwrap();
    let non_alphanumeric1 = !char1.is_ascii_alphanumeric();
    let non_alphanumeric2 = !char2.is_ascii_alphanumeric();
    let whitespace1 = non_alphanumeric1 && char1.is_ascii_whitespace();
    let whitespace2 = non_alphanumeric2 && char2.is_ascii_whitespace();
    let line_break1 = whitespace1 && char1.is_control();
    let line_break2 = whitespace2 && char2.is_control();
    let blank_line1 = line_break1 && (one.ends_with("\n\n") || one.ends_with("\n\r\n"));
    let blank_line2 = line_break2 && (two.starts_with("\n\n") || two.starts_with("\r\n\r\n"));

    if blank_line1 || blank_line2 {
        // Five points for blank lines.
        5
    } else if line_break1 || line_break2 {
        // Four points for line breaks.
        4
    } else if non_alphanumeric1 && !whitespace1 && whitespace2 {
        // Three points for end of sentences.
        3
    } else if whitespace1 || whitespace2 {
        // Two points for whitespace.
        2
    } else if non_alphanumeric1 || non_alphanumeric2 {
        // One point for non-alphanumeric.
        1
    } else {
        0
    }
}

// Reorder and merge like edit sections. Merge equalities. Any edit section can
// move as long as it doesn't cross an equality.
fn cleanup_merge(solution: &mut Solution) {
    let diffs = &mut solution.diffs;
    let common_prefix = if solution.utf8 {
        common_prefix
    } else {
        common_prefix_bytes
    };
    let common_suffix = if solution.utf8 {
        common_suffix
    } else {
        common_suffix_bytes
    };

    loop {
        if diffs.is_empty() {
            return;
        }

        diffs.push(Diff::Equal(
            solution.text1.substring(solution.text1.len..),
            solution.text2.substring(solution.text2.len..),
        )); // Add a dummy entry at the end.
        let mut pointer = 0;
        let mut count_delete = 0;
        let mut count_insert = 0;
        let mut text_delete = Range::empty();
        let mut text_insert = Range::empty();
        while let Some(&this_diff) = diffs.get(pointer) {
            match this_diff {
                Diff::Insert(text) => {
                    count_insert += 1;
                    if text_insert.is_empty() {
                        text_insert = text;
                    } else {
                        text_insert.len += text.len;
                    }
                }
                Diff::Delete(text) => {
                    count_delete += 1;
                    if text_delete.is_empty() {
                        text_delete = text;
                    } else {
                        text_delete.len += text.len;
                    }
                }
                Diff::Equal(text, _) => {
                    let count_both = count_delete + count_insert;
                    if count_both > 1 {
                        let both_types = count_delete != 0 && count_insert != 0;
                        // Delete the offending records.
                        diffs.splice(pointer - count_both..pointer, None);
                        pointer -= count_both;
                        if both_types {
                            // Factor out any common prefix.
                            let common_length = common_prefix(text_insert, text_delete);
                            if common_length != 0 {
                                if pointer > 0 {
                                    match &mut diffs[pointer - 1] {
                                        Diff::Equal(this_diff1, this_diff2) => {
                                            this_diff1.len += common_length;
                                            this_diff2.len += common_length;
                                        }
                                        _ => unreachable!(
                                            "previous diff should have been an equality"
                                        ),
                                    }
                                } else {
                                    diffs.insert(
                                        pointer,
                                        Diff::Equal(
                                            text_delete.substring(..common_length),
                                            text_insert.substring(..common_length),
                                        ),
                                    );
                                    pointer += 1;
                                }
                                text_insert = text_insert.substring(common_length..);
                                text_delete = text_delete.substring(common_length..);
                            }
                            // Factor out any common suffix.
                            let common_length = common_suffix(text_insert, text_delete);
                            if common_length != 0 {
                                diffs[pointer].grow_left(common_length);
                                text_insert.len -= common_length;
                                text_delete.len -= common_length;
                            }
                        }
                        // Insert the merged records.
                        if !text_delete.is_empty() {
                            diffs.insert(pointer, Diff::Delete(text_delete));
                            pointer += 1;
                        }
                        if !text_insert.is_empty() {
                            diffs.insert(pointer, Diff::Insert(text_insert));
                            pointer += 1;
                        }
                    } else if pointer > 0 {
                        if let Some(Diff::Equal(prev_equal1, prev_equal2)) =
                            diffs.get_mut(pointer - 1)
                        {
                            // Merge this equality with the previous one.
                            prev_equal1.len += text.len;
                            prev_equal2.len += text.len;
                            diffs.remove(pointer);
                            pointer -= 1;
                        }
                    }
                    count_insert = 0;
                    count_delete = 0;
                    text_delete = Range::empty();
                    text_insert = Range::empty();
                }
            }
            pointer += 1;
        }
        if diffs.last().unwrap().text().is_empty() {
            diffs.pop(); // Remove the dummy entry at the end.
        }

        // Second pass: look for single edits surrounded on both sides by equalities
        // which can be shifted sideways to eliminate an equality.
        // e.g: A<ins>BA</ins>C -> <ins>AB</ins>AC
        let mut changes = false;
        let mut pointer = 1;
        // Intentionally ignore the first and last element (don't need checking).
        while let Some(&next_diff) = diffs.get(pointer + 1) {
            let prev_diff = diffs[pointer - 1];
            let this_diff = diffs[pointer];
            if let (Diff::Equal(prev_diff, _), Diff::Equal(next_diff, _)) = (prev_diff, next_diff) {
                // This is a single edit surrounded by equalities.
                if this_diff.text().ends_with(prev_diff) {
                    // Shift the edit over the previous equality.
                    diffs[pointer].shift_left(prev_diff.len);
                    diffs[pointer + 1].grow_left(prev_diff.len);
                    diffs.remove(pointer - 1); // Delete prev_diff.
                    changes = true;
                } else if this_diff.text().starts_with(next_diff) {
                    // Shift the edit over the next equality.
                    diffs[pointer - 1].grow_right(next_diff.len);
                    diffs[pointer].shift_right(next_diff.len);
                    diffs.remove(pointer + 1); // Delete next_diff.
                    changes = true;
                }
            }
            pointer += 1;
        }
        // If shifts were made, the diff needs reordering and another shift sweep.
        if !changes {
            return;
        }
    }
}

impl Debug for Chunk<'_> {
    fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
        let (name, text) = match *self {
            Chunk::Equal(text) => ("Equal", text),
            Chunk::Delete(text) => ("Delete", text),
            Chunk::Insert(text) => ("Insert", text),
        };
        write!(formatter, "{}({:?})", name, text)
    }
}

impl Debug for Diff<'_, '_> {
    fn fmt(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
        let (name, bytes) = match *self {
            Diff::Equal(range, _) => ("Equal", bytes(range)),
            Diff::Delete(range) => ("Delete", bytes(range)),
            Diff::Insert(range) => ("Insert", bytes(range)),
        };
        let text = String::from_utf8_lossy(bytes);
        write!(formatter, "{}({:?})", name, text)
    }
}

impl<'a> From<Diff<'a, 'a>> for Chunk<'a> {
    fn from(diff: Diff<'a, 'a>) -> Self {
        match diff {
            Diff::Equal(range, _) => Chunk::Equal(str(range)),
            Diff::Delete(range) => Chunk::Delete(str(range)),
            Diff::Insert(range) => Chunk::Insert(str(range)),
        }
    }
}