summaryrefslogtreecommitdiff
path: root/deps/icu-small/source/common/edits.cpp
blob: 00a8d601a1cc80fbfcc46b7e1c677803e41511b8 (plain)
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
// © 2017 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html

// edits.cpp
// created: 2017feb08 Markus W. Scherer

#include "unicode/edits.h"
#include "unicode/unistr.h"
#include "unicode/utypes.h"
#include "cmemory.h"
#include "uassert.h"
#include "util.h"

U_NAMESPACE_BEGIN

namespace {

// 0000uuuuuuuuuuuu records u+1 unchanged text units.
const int32_t MAX_UNCHANGED_LENGTH = 0x1000;
const int32_t MAX_UNCHANGED = MAX_UNCHANGED_LENGTH - 1;

// 0mmmnnnccccccccc with m=1..6 records ccc+1 replacements of m:n text units.
const int32_t MAX_SHORT_CHANGE_OLD_LENGTH = 6;
const int32_t MAX_SHORT_CHANGE_NEW_LENGTH = 7;
const int32_t SHORT_CHANGE_NUM_MASK = 0x1ff;
const int32_t MAX_SHORT_CHANGE = 0x6fff;

// 0111mmmmmmnnnnnn records a replacement of m text units with n.
// m or n = 61: actual length follows in the next edits array unit.
// m or n = 62..63: actual length follows in the next two edits array units.
// Bit 30 of the actual length is in the head unit.
// Trailing units have bit 15 set.
const int32_t LENGTH_IN_1TRAIL = 61;
const int32_t LENGTH_IN_2TRAIL = 62;

}  // namespace

void Edits::releaseArray() U_NOEXCEPT {
    if (array != stackArray) {
        uprv_free(array);
    }
}

Edits &Edits::copyArray(const Edits &other) {
    if (U_FAILURE(errorCode_)) {
        length = delta = numChanges = 0;
        return *this;
    }
    if (length > capacity) {
        uint16_t *newArray = (uint16_t *)uprv_malloc((size_t)length * 2);
        if (newArray == nullptr) {
            length = delta = numChanges = 0;
            errorCode_ = U_MEMORY_ALLOCATION_ERROR;
            return *this;
        }
        releaseArray();
        array = newArray;
        capacity = length;
    }
    if (length > 0) {
        uprv_memcpy(array, other.array, (size_t)length * 2);
    }
    return *this;
}

Edits &Edits::moveArray(Edits &src) U_NOEXCEPT {
    if (U_FAILURE(errorCode_)) {
        length = delta = numChanges = 0;
        return *this;
    }
    releaseArray();
    if (length > STACK_CAPACITY) {
        array = src.array;
        capacity = src.capacity;
        src.array = src.stackArray;
        src.capacity = STACK_CAPACITY;
        src.reset();
        return *this;
    }
    array = stackArray;
    capacity = STACK_CAPACITY;
    if (length > 0) {
        uprv_memcpy(array, src.array, (size_t)length * 2);
    }
    return *this;
}

Edits &Edits::operator=(const Edits &other) {
    length = other.length;
    delta = other.delta;
    numChanges = other.numChanges;
    errorCode_ = other.errorCode_;
    return copyArray(other);
}

Edits &Edits::operator=(Edits &&src) U_NOEXCEPT {
    length = src.length;
    delta = src.delta;
    numChanges = src.numChanges;
    errorCode_ = src.errorCode_;
    return moveArray(src);
}

Edits::~Edits() {
    releaseArray();
}

void Edits::reset() U_NOEXCEPT {
    length = delta = numChanges = 0;
    errorCode_ = U_ZERO_ERROR;
}

void Edits::addUnchanged(int32_t unchangedLength) {
    if(U_FAILURE(errorCode_) || unchangedLength == 0) { return; }
    if(unchangedLength < 0) {
        errorCode_ = U_ILLEGAL_ARGUMENT_ERROR;
        return;
    }
    // Merge into previous unchanged-text record, if any.
    int32_t last = lastUnit();
    if(last < MAX_UNCHANGED) {
        int32_t remaining = MAX_UNCHANGED - last;
        if (remaining >= unchangedLength) {
            setLastUnit(last + unchangedLength);
            return;
        }
        setLastUnit(MAX_UNCHANGED);
        unchangedLength -= remaining;
    }
    // Split large lengths into multiple units.
    while(unchangedLength >= MAX_UNCHANGED_LENGTH) {
        append(MAX_UNCHANGED);
        unchangedLength -= MAX_UNCHANGED_LENGTH;
    }
    // Write a small (remaining) length.
    if(unchangedLength > 0) {
        append(unchangedLength - 1);
    }
}

void Edits::addReplace(int32_t oldLength, int32_t newLength) {
    if(U_FAILURE(errorCode_)) { return; }
    if(oldLength < 0 || newLength < 0) {
        errorCode_ = U_ILLEGAL_ARGUMENT_ERROR;
        return;
    }
    if (oldLength == 0 && newLength == 0) {
        return;
    }
    ++numChanges;
    int32_t newDelta = newLength - oldLength;
    if (newDelta != 0) {
        if ((newDelta > 0 && delta >= 0 && newDelta > (INT32_MAX - delta)) ||
                (newDelta < 0 && delta < 0 && newDelta < (INT32_MIN - delta))) {
            // Integer overflow or underflow.
            errorCode_ = U_INDEX_OUTOFBOUNDS_ERROR;
            return;
        }
        delta += newDelta;
    }

    if(0 < oldLength && oldLength <= MAX_SHORT_CHANGE_OLD_LENGTH &&
            newLength <= MAX_SHORT_CHANGE_NEW_LENGTH) {
        // Merge into previous same-lengths short-replacement record, if any.
        int32_t u = (oldLength << 12) | (newLength << 9);
        int32_t last = lastUnit();
        if(MAX_UNCHANGED < last && last < MAX_SHORT_CHANGE &&
                (last & ~SHORT_CHANGE_NUM_MASK) == u &&
                (last & SHORT_CHANGE_NUM_MASK) < SHORT_CHANGE_NUM_MASK) {
            setLastUnit(last + 1);
            return;
        }
        append(u);
        return;
    }

    int32_t head = 0x7000;
    if (oldLength < LENGTH_IN_1TRAIL && newLength < LENGTH_IN_1TRAIL) {
        head |= oldLength << 6;
        head |= newLength;
        append(head);
    } else if ((capacity - length) >= 5 || growArray()) {
        int32_t limit = length + 1;
        if(oldLength < LENGTH_IN_1TRAIL) {
            head |= oldLength << 6;
        } else if(oldLength <= 0x7fff) {
            head |= LENGTH_IN_1TRAIL << 6;
            array[limit++] = (uint16_t)(0x8000 | oldLength);
        } else {
            head |= (LENGTH_IN_2TRAIL + (oldLength >> 30)) << 6;
            array[limit++] = (uint16_t)(0x8000 | (oldLength >> 15));
            array[limit++] = (uint16_t)(0x8000 | oldLength);
        }
        if(newLength < LENGTH_IN_1TRAIL) {
            head |= newLength;
        } else if(newLength <= 0x7fff) {
            head |= LENGTH_IN_1TRAIL;
            array[limit++] = (uint16_t)(0x8000 | newLength);
        } else {
            head |= LENGTH_IN_2TRAIL + (newLength >> 30);
            array[limit++] = (uint16_t)(0x8000 | (newLength >> 15));
            array[limit++] = (uint16_t)(0x8000 | newLength);
        }
        array[length] = (uint16_t)head;
        length = limit;
    }
}

void Edits::append(int32_t r) {
    if(length < capacity || growArray()) {
        array[length++] = (uint16_t)r;
    }
}

UBool Edits::growArray() {
    int32_t newCapacity;
    if (array == stackArray) {
        newCapacity = 2000;
    } else if (capacity == INT32_MAX) {
        // Not U_BUFFER_OVERFLOW_ERROR because that could be confused on a string transform API
        // with a result-string-buffer overflow.
        errorCode_ = U_INDEX_OUTOFBOUNDS_ERROR;
        return FALSE;
    } else if (capacity >= (INT32_MAX / 2)) {
        newCapacity = INT32_MAX;
    } else {
        newCapacity = 2 * capacity;
    }
    // Grow by at least 5 units so that a maximal change record will fit.
    if ((newCapacity - capacity) < 5) {
        errorCode_ = U_INDEX_OUTOFBOUNDS_ERROR;
        return FALSE;
    }
    uint16_t *newArray = (uint16_t *)uprv_malloc((size_t)newCapacity * 2);
    if (newArray == NULL) {
        errorCode_ = U_MEMORY_ALLOCATION_ERROR;
        return FALSE;
    }
    uprv_memcpy(newArray, array, (size_t)length * 2);
    releaseArray();
    array = newArray;
    capacity = newCapacity;
    return TRUE;
}

UBool Edits::copyErrorTo(UErrorCode &outErrorCode) {
    if (U_FAILURE(outErrorCode)) { return TRUE; }
    if (U_SUCCESS(errorCode_)) { return FALSE; }
    outErrorCode = errorCode_;
    return TRUE;
}

Edits &Edits::mergeAndAppend(const Edits &ab, const Edits &bc, UErrorCode &errorCode) {
    if (copyErrorTo(errorCode)) { return *this; }
    // Picture string a --(Edits ab)--> string b --(Edits bc)--> string c.
    // Parallel iteration over both Edits.
    Iterator abIter = ab.getFineIterator();
    Iterator bcIter = bc.getFineIterator();
    UBool abHasNext = TRUE, bcHasNext = TRUE;
    // Copy iterator state into local variables, so that we can modify and subdivide spans.
    // ab old & new length, bc old & new length
    int32_t aLength = 0, ab_bLength = 0, bc_bLength = 0, cLength = 0;
    // When we have different-intermediate-length changes, we accumulate a larger change.
    int32_t pending_aLength = 0, pending_cLength = 0;
    for (;;) {
        // At this point, for each of the two iterators:
        // Either we are done with the locally cached current edit,
        // and its intermediate-string length has been reset,
        // or we will continue to work with a truncated remainder of this edit.
        //
        // If the current edit is done, and the iterator has not yet reached the end,
        // then we fetch the next edit. This is true for at least one of the iterators.
        //
        // Normally it does not matter whether we fetch from ab and then bc or vice versa.
        // However, the result is observably different when
        // ab deletions meet bc insertions at the same intermediate-string index.
        // Some users expect the bc insertions to come first, so we fetch from bc first.
        if (bc_bLength == 0) {
            if (bcHasNext && (bcHasNext = bcIter.next(errorCode)) != 0) {
                bc_bLength = bcIter.oldLength();
                cLength = bcIter.newLength();
                if (bc_bLength == 0) {
                    // insertion
                    if (ab_bLength == 0 || !abIter.hasChange()) {
                        addReplace(pending_aLength, pending_cLength + cLength);
                        pending_aLength = pending_cLength = 0;
                    } else {
                        pending_cLength += cLength;
                    }
                    continue;
                }
            }
            // else see if the other iterator is done, too.
        }
        if (ab_bLength == 0) {
            if (abHasNext && (abHasNext = abIter.next(errorCode)) != 0) {
                aLength = abIter.oldLength();
                ab_bLength = abIter.newLength();
                if (ab_bLength == 0) {
                    // deletion
                    if (bc_bLength == bcIter.oldLength() || !bcIter.hasChange()) {
                        addReplace(pending_aLength + aLength, pending_cLength);
                        pending_aLength = pending_cLength = 0;
                    } else {
                        pending_aLength += aLength;
                    }
                    continue;
                }
            } else if (bc_bLength == 0) {
                // Both iterators are done at the same time:
                // The intermediate-string lengths match.
                break;
            } else {
                // The ab output string is shorter than the bc input string.
                if (!copyErrorTo(errorCode)) {
                    errorCode = U_ILLEGAL_ARGUMENT_ERROR;
                }
                return *this;
            }
        }
        if (bc_bLength == 0) {
            // The bc input string is shorter than the ab output string.
            if (!copyErrorTo(errorCode)) {
                errorCode = U_ILLEGAL_ARGUMENT_ERROR;
            }
            return *this;
        }
        //  Done fetching: ab_bLength > 0 && bc_bLength > 0

        // The current state has two parts:
        // - Past: We accumulate a longer ac edit in the "pending" variables.
        // - Current: We have copies of the current ab/bc edits in local variables.
        //   At least one side is newly fetched.
        //   One side might be a truncated remainder of an edit we fetched earlier.

        if (!abIter.hasChange() && !bcIter.hasChange()) {
            // An unchanged span all the way from string a to string c.
            if (pending_aLength != 0 || pending_cLength != 0) {
                addReplace(pending_aLength, pending_cLength);
                pending_aLength = pending_cLength = 0;
            }
            int32_t unchangedLength = aLength <= cLength ? aLength : cLength;
            addUnchanged(unchangedLength);
            ab_bLength = aLength -= unchangedLength;
            bc_bLength = cLength -= unchangedLength;
            // At least one of the unchanged spans is now empty.
            continue;
        }
        if (!abIter.hasChange() && bcIter.hasChange()) {
            // Unchanged a->b but changed b->c.
            if (ab_bLength >= bc_bLength) {
                // Split the longer unchanged span into change + remainder.
                addReplace(pending_aLength + bc_bLength, pending_cLength + cLength);
                pending_aLength = pending_cLength = 0;
                aLength = ab_bLength -= bc_bLength;
                bc_bLength = 0;
                continue;
            }
            // Handle the shorter unchanged span below like a change.
        } else if (abIter.hasChange() && !bcIter.hasChange()) {
            // Changed a->b and then unchanged b->c.
            if (ab_bLength <= bc_bLength) {
                // Split the longer unchanged span into change + remainder.
                addReplace(pending_aLength + aLength, pending_cLength + ab_bLength);
                pending_aLength = pending_cLength = 0;
                cLength = bc_bLength -= ab_bLength;
                ab_bLength = 0;
                continue;
            }
            // Handle the shorter unchanged span below like a change.
        } else {  // both abIter.hasChange() && bcIter.hasChange()
            if (ab_bLength == bc_bLength) {
                // Changes on both sides up to the same position. Emit & reset.
                addReplace(pending_aLength + aLength, pending_cLength + cLength);
                pending_aLength = pending_cLength = 0;
                ab_bLength = bc_bLength = 0;
                continue;
            }
        }
        // Accumulate the a->c change, reset the shorter side,
        // keep a remainder of the longer one.
        pending_aLength += aLength;
        pending_cLength += cLength;
        if (ab_bLength < bc_bLength) {
            bc_bLength -= ab_bLength;
            cLength = ab_bLength = 0;
        } else {  // ab_bLength > bc_bLength
            ab_bLength -= bc_bLength;
            aLength = bc_bLength = 0;
        }
    }
    if (pending_aLength != 0 || pending_cLength != 0) {
        addReplace(pending_aLength, pending_cLength);
    }
    copyErrorTo(errorCode);
    return *this;
}

Edits::Iterator::Iterator(const uint16_t *a, int32_t len, UBool oc, UBool crs) :
        array(a), index(0), length(len), remaining(0),
        onlyChanges_(oc), coarse(crs),
        dir(0), changed(FALSE), oldLength_(0), newLength_(0),
        srcIndex(0), replIndex(0), destIndex(0) {}

int32_t Edits::Iterator::readLength(int32_t head) {
    if (head < LENGTH_IN_1TRAIL) {
        return head;
    } else if (head < LENGTH_IN_2TRAIL) {
        U_ASSERT(index < length);
        U_ASSERT(array[index] >= 0x8000);
        return array[index++] & 0x7fff;
    } else {
        U_ASSERT((index + 2) <= length);
        U_ASSERT(array[index] >= 0x8000);
        U_ASSERT(array[index + 1] >= 0x8000);
        int32_t len = ((head & 1) << 30) |
                ((int32_t)(array[index] & 0x7fff) << 15) |
                (array[index + 1] & 0x7fff);
        index += 2;
        return len;
    }
}

void Edits::Iterator::updateNextIndexes() {
    srcIndex += oldLength_;
    if (changed) {
        replIndex += newLength_;
    }
    destIndex += newLength_;
}

void Edits::Iterator::updatePreviousIndexes() {
    srcIndex -= oldLength_;
    if (changed) {
        replIndex -= newLength_;
    }
    destIndex -= newLength_;
}

UBool Edits::Iterator::noNext() {
    // No change before or beyond the string.
    dir = 0;
    changed = FALSE;
    oldLength_ = newLength_ = 0;
    return FALSE;
}

UBool Edits::Iterator::next(UBool onlyChanges, UErrorCode &errorCode) {
    // Forward iteration: Update the string indexes to the limit of the current span,
    // and post-increment-read array units to assemble a new span.
    // Leaves the array index one after the last unit of that span.
    if (U_FAILURE(errorCode)) { return FALSE; }
    // We have an errorCode in case we need to start guarding against integer overflows.
    // It is also convenient for caller loops if we bail out when an error was set elsewhere.
    if (dir > 0) {
        updateNextIndexes();
    } else {
        if (dir < 0) {
            // Turn around from previous() to next().
            // Post-increment-read the same span again.
            if (remaining > 0) {
                // Fine-grained iterator:
                // Stay on the current one of a sequence of compressed changes.
                ++index;  // next() rests on the index after the sequence unit.
                dir = 1;
                return TRUE;
            }
        }
        dir = 1;
    }
    if (remaining >= 1) {
        // Fine-grained iterator: Continue a sequence of compressed changes.
        if (remaining > 1) {
            --remaining;
            return TRUE;
        }
        remaining = 0;
    }
    if (index >= length) {
        return noNext();
    }
    int32_t u = array[index++];
    if (u <= MAX_UNCHANGED) {
        // Combine adjacent unchanged ranges.
        changed = FALSE;
        oldLength_ = u + 1;
        while (index < length && (u = array[index]) <= MAX_UNCHANGED) {
            ++index;
            oldLength_ += u + 1;
        }
        newLength_ = oldLength_;
        if (onlyChanges) {
            updateNextIndexes();
            if (index >= length) {
                return noNext();
            }
            // already fetched u > MAX_UNCHANGED at index
            ++index;
        } else {
            return TRUE;
        }
    }
    changed = TRUE;
    if (u <= MAX_SHORT_CHANGE) {
        int32_t oldLen = u >> 12;
        int32_t newLen = (u >> 9) & MAX_SHORT_CHANGE_NEW_LENGTH;
        int32_t num = (u & SHORT_CHANGE_NUM_MASK) + 1;
        if (coarse) {
            oldLength_ = num * oldLen;
            newLength_ = num * newLen;
        } else {
            // Split a sequence of changes that was compressed into one unit.
            oldLength_ = oldLen;
            newLength_ = newLen;
            if (num > 1) {
                remaining = num;  // This is the first of two or more changes.
            }
            return TRUE;
        }
    } else {
        U_ASSERT(u <= 0x7fff);
        oldLength_ = readLength((u >> 6) & 0x3f);
        newLength_ = readLength(u & 0x3f);
        if (!coarse) {
            return TRUE;
        }
    }
    // Combine adjacent changes.
    while (index < length && (u = array[index]) > MAX_UNCHANGED) {
        ++index;
        if (u <= MAX_SHORT_CHANGE) {
            int32_t num = (u & SHORT_CHANGE_NUM_MASK) + 1;
            oldLength_ += (u >> 12) * num;
            newLength_ += ((u >> 9) & MAX_SHORT_CHANGE_NEW_LENGTH) * num;
        } else {
            U_ASSERT(u <= 0x7fff);
            oldLength_ += readLength((u >> 6) & 0x3f);
            newLength_ += readLength(u & 0x3f);
        }
    }
    return TRUE;
}

UBool Edits::Iterator::previous(UErrorCode &errorCode) {
    // Backward iteration: Pre-decrement-read array units to assemble a new span,
    // then update the string indexes to the start of that span.
    // Leaves the array index on the head unit of that span.
    if (U_FAILURE(errorCode)) { return FALSE; }
    // We have an errorCode in case we need to start guarding against integer overflows.
    // It is also convenient for caller loops if we bail out when an error was set elsewhere.
    if (dir >= 0) {
        if (dir > 0) {
            // Turn around from next() to previous().
            // Set the string indexes to the span limit and
            // pre-decrement-read the same span again.
            if (remaining > 0) {
                // Fine-grained iterator:
                // Stay on the current one of a sequence of compressed changes.
                --index;  // previous() rests on the sequence unit.
                dir = -1;
                return TRUE;
            }
            updateNextIndexes();
        }
        dir = -1;
    }
    if (remaining > 0) {
        // Fine-grained iterator: Continue a sequence of compressed changes.
        int32_t u = array[index];
        U_ASSERT(MAX_UNCHANGED < u && u <= MAX_SHORT_CHANGE);
        if (remaining <= (u & SHORT_CHANGE_NUM_MASK)) {
            ++remaining;
            updatePreviousIndexes();
            return TRUE;
        }
        remaining = 0;
    }
    if (index <= 0) {
        return noNext();
    }
    int32_t u = array[--index];
    if (u <= MAX_UNCHANGED) {
        // Combine adjacent unchanged ranges.
        changed = FALSE;
        oldLength_ = u + 1;
        while (index > 0 && (u = array[index - 1]) <= MAX_UNCHANGED) {
            --index;
            oldLength_ += u + 1;
        }
        newLength_ = oldLength_;
        // No need to handle onlyChanges as long as previous() is called only from findIndex().
        updatePreviousIndexes();
        return TRUE;
    }
    changed = TRUE;
    if (u <= MAX_SHORT_CHANGE) {
        int32_t oldLen = u >> 12;
        int32_t newLen = (u >> 9) & MAX_SHORT_CHANGE_NEW_LENGTH;
        int32_t num = (u & SHORT_CHANGE_NUM_MASK) + 1;
        if (coarse) {
            oldLength_ = num * oldLen;
            newLength_ = num * newLen;
        } else {
            // Split a sequence of changes that was compressed into one unit.
            oldLength_ = oldLen;
            newLength_ = newLen;
            if (num > 1) {
                remaining = 1;  // This is the last of two or more changes.
            }
            updatePreviousIndexes();
            return TRUE;
        }
    } else {
        if (u <= 0x7fff) {
            // The change is encoded in u alone.
            oldLength_ = readLength((u >> 6) & 0x3f);
            newLength_ = readLength(u & 0x3f);
        } else {
            // Back up to the head of the change, read the lengths,
            // and reset the index to the head again.
            U_ASSERT(index > 0);
            while ((u = array[--index]) > 0x7fff) {}
            U_ASSERT(u > MAX_SHORT_CHANGE);
            int32_t headIndex = index++;
            oldLength_ = readLength((u >> 6) & 0x3f);
            newLength_ = readLength(u & 0x3f);
            index = headIndex;
        }
        if (!coarse) {
            updatePreviousIndexes();
            return TRUE;
        }
    }
    // Combine adjacent changes.
    while (index > 0 && (u = array[index - 1]) > MAX_UNCHANGED) {
        --index;
        if (u <= MAX_SHORT_CHANGE) {
            int32_t num = (u & SHORT_CHANGE_NUM_MASK) + 1;
            oldLength_ += (u >> 12) * num;
            newLength_ += ((u >> 9) & MAX_SHORT_CHANGE_NEW_LENGTH) * num;
        } else if (u <= 0x7fff) {
            // Read the lengths, and reset the index to the head again.
            int32_t headIndex = index++;
            oldLength_ += readLength((u >> 6) & 0x3f);
            newLength_ += readLength(u & 0x3f);
            index = headIndex;
        }
    }
    updatePreviousIndexes();
    return TRUE;
}

int32_t Edits::Iterator::findIndex(int32_t i, UBool findSource, UErrorCode &errorCode) {
    if (U_FAILURE(errorCode) || i < 0) { return -1; }
    int32_t spanStart, spanLength;
    if (findSource) {  // find source index
        spanStart = srcIndex;
        spanLength = oldLength_;
    } else {  // find destination index
        spanStart = destIndex;
        spanLength = newLength_;
    }
    if (i < spanStart) {
        if (i >= (spanStart / 2)) {
            // Search backwards.
            for (;;) {
                UBool hasPrevious = previous(errorCode);
                U_ASSERT(hasPrevious);  // because i>=0 and the first span starts at 0
                (void)hasPrevious;  // avoid unused-variable warning
                spanStart = findSource ? srcIndex : destIndex;
                if (i >= spanStart) {
                    // The index is in the current span.
                    return 0;
                }
                if (remaining > 0) {
                    // Is the index in one of the remaining compressed edits?
                    // spanStart is the start of the current span, first of the remaining ones.
                    spanLength = findSource ? oldLength_ : newLength_;
                    int32_t u = array[index];
                    U_ASSERT(MAX_UNCHANGED < u && u <= MAX_SHORT_CHANGE);
                    int32_t num = (u & SHORT_CHANGE_NUM_MASK) + 1 - remaining;
                    int32_t len = num * spanLength;
                    if (i >= (spanStart - len)) {
                        int32_t n = ((spanStart - i - 1) / spanLength) + 1;
                        // 1 <= n <= num
                        srcIndex -= n * oldLength_;
                        replIndex -= n * newLength_;
                        destIndex -= n * newLength_;
                        remaining += n;
                        return 0;
                    }
                    // Skip all of these edits at once.
                    srcIndex -= num * oldLength_;
                    replIndex -= num * newLength_;
                    destIndex -= num * newLength_;
                    remaining = 0;
                }
            }
        }
        // Reset the iterator to the start.
        dir = 0;
        index = remaining = oldLength_ = newLength_ = srcIndex = replIndex = destIndex = 0;
    } else if (i < (spanStart + spanLength)) {
        // The index is in the current span.
        return 0;
    }
    while (next(FALSE, errorCode)) {
        if (findSource) {
            spanStart = srcIndex;
            spanLength = oldLength_;
        } else {
            spanStart = destIndex;
            spanLength = newLength_;
        }
        if (i < (spanStart + spanLength)) {
            // The index is in the current span.
            return 0;
        }
        if (remaining > 1) {
            // Is the index in one of the remaining compressed edits?
            // spanStart is the start of the current span, first of the remaining ones.
            int32_t len = remaining * spanLength;
            if (i < (spanStart + len)) {
                int32_t n = (i - spanStart) / spanLength;  // 1 <= n <= remaining - 1
                srcIndex += n * oldLength_;
                replIndex += n * newLength_;
                destIndex += n * newLength_;
                remaining -= n;
                return 0;
            }
            // Make next() skip all of these edits at once.
            oldLength_ *= remaining;
            newLength_ *= remaining;
            remaining = 0;
        }
    }
    return 1;
}

int32_t Edits::Iterator::destinationIndexFromSourceIndex(int32_t i, UErrorCode &errorCode) {
    int32_t where = findIndex(i, TRUE, errorCode);
    if (where < 0) {
        // Error or before the string.
        return 0;
    }
    if (where > 0 || i == srcIndex) {
        // At or after string length, or at start of the found span.
        return destIndex;
    }
    if (changed) {
        // In a change span, map to its end.
        return destIndex + newLength_;
    } else {
        // In an unchanged span, offset 1:1 within it.
        return destIndex + (i - srcIndex);
    }
}

int32_t Edits::Iterator::sourceIndexFromDestinationIndex(int32_t i, UErrorCode &errorCode) {
    int32_t where = findIndex(i, FALSE, errorCode);
    if (where < 0) {
        // Error or before the string.
        return 0;
    }
    if (where > 0 || i == destIndex) {
        // At or after string length, or at start of the found span.
        return srcIndex;
    }
    if (changed) {
        // In a change span, map to its end.
        return srcIndex + oldLength_;
    } else {
        // In an unchanged span, offset within it.
        return srcIndex + (i - destIndex);
    }
}

UnicodeString& Edits::Iterator::toString(UnicodeString& sb) const {
    sb.append(u"{ src[", -1);
    ICU_Utility::appendNumber(sb, srcIndex);
    sb.append(u"..", -1);
    ICU_Utility::appendNumber(sb, srcIndex + oldLength_);
    if (changed) {
        sb.append(u"] ⇝ dest[", -1);
    } else {
        sb.append(u"] ≡ dest[", -1);
    }
    ICU_Utility::appendNumber(sb, destIndex);
    sb.append(u"..", -1);
    ICU_Utility::appendNumber(sb, destIndex + newLength_);
    if (changed) {
        sb.append(u"], repl[", -1);
        ICU_Utility::appendNumber(sb, replIndex);
        sb.append(u"..", -1);
        ICU_Utility::appendNumber(sb, replIndex + newLength_);
        sb.append(u"] }", -1);
    } else {
        sb.append(u"] (no-change) }", -1);
    }
    return sb;
}

U_NAMESPACE_END