File: case_map.cc

package info (click to toggle)
chromium 138.0.7204.183-1
  • links: PTS, VCS
  • area: main
  • in suites: trixie
  • size: 6,071,908 kB
  • sloc: cpp: 34,937,088; ansic: 7,176,967; javascript: 4,110,704; python: 1,419,953; asm: 946,768; xml: 739,971; pascal: 187,324; sh: 89,623; perl: 88,663; objc: 79,944; sql: 50,304; cs: 41,786; fortran: 24,137; makefile: 21,806; php: 13,980; tcl: 13,166; yacc: 8,925; ruby: 7,485; awk: 3,720; lisp: 3,096; lex: 1,327; ada: 727; jsp: 228; sed: 36
file content (436 lines) | stat: -rw-r--r-- 15,237 bytes parent folder | download | duplicates (2)
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
// Copyright 2019 The Chromium Authors
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.

#include "third_party/blink/renderer/platform/wtf/text/case_map.h"

#include <unicode/casemap.h>

#include "base/notreached.h"
#include "third_party/blink/renderer/platform/wtf/text/atomic_string.h"
#include "third_party/blink/renderer/platform/wtf/text/character_names.h"
#include "third_party/blink/renderer/platform/wtf/text/string_buffer.h"
#include "third_party/blink/renderer/platform/wtf/text/string_impl.h"
#include "third_party/blink/renderer/platform/wtf/text/string_view.h"
#include "third_party/blink/renderer/platform/wtf/text/text_offset_map.h"
#include "third_party/blink/renderer/platform/wtf/text/unicode.h"
#include "third_party/blink/renderer/platform/wtf/text/wtf_string.h"

namespace WTF {

namespace {

// `lang` - Language code.  Now it is one of "tr", "az", "el", and "lt".
inline bool LocaleIdMatchesLang(const AtomicString& locale_id,
                                const StringView lang) {
  const wtf_size_t lang_length = lang.length();
  CHECK(lang_length == 2u || lang_length == 3u);
  const StringImpl* locale_id_impl = locale_id.Impl();
  if (!locale_id_impl || !locale_id_impl->StartsWithIgnoringASCIICase(lang)) {
    return false;
  }
  if (locale_id_impl->length() == lang_length) {
    return true;
  }
  const UChar maybe_delimiter = (*locale_id_impl)[lang_length];
  return maybe_delimiter == '-' || maybe_delimiter == '_' ||
         maybe_delimiter == '@';
}

enum class CaseMapType { kLower, kUpper, kTitle };

icu::Edits DecreaseFirstEditLength(const icu::Edits& edits) {
  icu::Edits new_edits;
  UErrorCode error = U_ZERO_ERROR;
  auto edit = edits.getFineIterator();

  edit.next(error);
  int32_t new_length = edit.oldLength() - 1;
  if (new_length > 0) {
    new_edits.addUnchanged(new_length);
  }

  while (edit.next(error)) {
    if (edit.hasChange()) {
      new_edits.addReplace(edit.oldLength(), edit.newLength());
    } else {
      new_edits.addUnchanged(edit.oldLength());
    }
  }
  DCHECK(U_SUCCESS(error));
  return new_edits;
}

scoped_refptr<StringImpl> CaseConvert(CaseMapType type,
                                      StringImpl* source,
                                      const char* locale,
                                      TextOffsetMap* offset_map = nullptr,
                                      UChar previous_character = 0) {
  DCHECK(source);
  CHECK_LE(source->length(),
           static_cast<wtf_size_t>(std::numeric_limits<int32_t>::max()));

  scoped_refptr<StringImpl> upconverted = source->UpconvertedString();
  const base::span<const UChar> source16 = upconverted->Span16();

  base::span<UChar> data16;
  scoped_refptr<StringImpl> output =
      StringImpl::CreateUninitialized(source16.size(), data16);
  wtf_size_t target_length = 0;
  while (true) {
    UErrorCode status = U_ZERO_ERROR;
    icu::Edits edits;
    switch (type) {
      case CaseMapType::kLower:
        target_length = icu::CaseMap::toLower(
            locale, /* options */ 0,
            reinterpret_cast<const char16_t*>(source16.data()), source16.size(),
            reinterpret_cast<char16_t*>(data16.data()), data16.size(), &edits,
            status);
        break;
      case CaseMapType::kUpper:
        target_length = icu::CaseMap::toUpper(
            locale, /* options */ 0,
            reinterpret_cast<const char16_t*>(source16.data()), source16.size(),
            reinterpret_cast<char16_t*>(data16.data()), data16.size(), &edits,
            status);
        break;
      case CaseMapType::kTitle: {
        unsigned source_length = source16.size();
        StringBuffer<UChar> string_buffer(source_length + 1);
        base::span<UChar> string_with_previous = string_buffer.Span();
        bool is_previous_alpha = u_isalpha(previous_character);
        // Use an 'A' as a previous character which is already capitalized to
        // make sure the titlecasing with previous character.
        string_with_previous[0] = is_previous_alpha ? u'A' : kSpaceCharacter;
        string_with_previous.subspan(1u).copy_from(source16);

        // Add a space of the previous character at the start.
        unsigned data_with_previous_length =
            (target_length ? target_length : source_length) + 1;
        StringBuffer<UChar> data_buffer(data_with_previous_length);
        base::span<UChar> data_with_previous = data_buffer.Span();

        target_length = icu::CaseMap::toTitle(
            locale, U_TITLECASE_NO_LOWERCASE, /* iter */ nullptr,
            reinterpret_cast<const char16_t*>(string_with_previous.data()),
            string_with_previous.size(),
            reinterpret_cast<char16_t*>(data_with_previous.data()),
            data_with_previous.size(), &edits, status);

        if (U_FAILURE(status)) {
          break;
        }

        // Remove the space of the previous character at the start.
        --target_length;
        // Copy the result without the previous character.
        data16.copy_from(data_with_previous.subspan(1u));
        // Since the text included the previous character, decrease length of
        // the first edit entry.
        edits = DecreaseFirstEditLength(edits);
        break;
      }
    }
    if (U_SUCCESS(status)) {
      if (!edits.hasChanges())
        return source;

      if (offset_map)
        offset_map->Append(edits);

      if (source16.size() == target_length) {
        return output;
      }
      return output->Substring(0, target_length);
    }

    // Expand the buffer and retry if the target is longer.
    if (status == U_BUFFER_OVERFLOW_ERROR) {
      output = StringImpl::CreateUninitialized(target_length, data16);
      continue;
    }

    NOTREACHED();
  }
}

}  // namespace

const char* CaseMap::Locale::turkic_or_azeri_ = "tr";
const char* CaseMap::Locale::greek_ = "el";
const char* CaseMap::Locale::lithuanian_ = "lt";
const char* CaseMap::Locale::dutch_ = "nl";

CaseMap::Locale::Locale(const AtomicString& locale) {
  // Use the more optimized code path most of the time.
  //
  // Only Turkic (tr and az) languages and Lithuanian requires
  // locale-specific lowercasing rules. Even though CLDR has el-Lower,
  // it's identical to the locale-agnostic lowercasing. Context-dependent
  // handling of Greek capital sigma is built into the common lowercasing
  // function in ICU.
  //
  // Only Turkic (tr and az) languages, Greek and Lithuanian require
  // locale-specific uppercasing rules.
  //
  // Only Dutch language requires locale-specific titlecasing rules.
  if (LocaleIdMatchesLang(locale, "tr") || LocaleIdMatchesLang(locale, "az"))
      [[unlikely]] {
    case_map_locale_ = turkic_or_azeri_;
  } else if (LocaleIdMatchesLang(locale, "el")) [[unlikely]] {
    case_map_locale_ = greek_;
  } else if (LocaleIdMatchesLang(locale, "lt")) [[unlikely]] {
    case_map_locale_ = lithuanian_;
  } else if (LocaleIdMatchesLang(locale, "nl")) [[unlikely]] {
    case_map_locale_ = dutch_;
  } else {
    case_map_locale_ = nullptr;
  }
}

scoped_refptr<StringImpl> CaseMap::TryFastToLowerInvariant(StringImpl* source) {
  DCHECK(source);

  // Note: This is a hot function in the Dromaeo benchmark, specifically the
  // no-op code path up through the first 'return' statement.

  // First scan the string for uppercase and non-ASCII characters:
  if (source->Is8Bit()) {
    const base::span<const LChar> source8 = source->Span8();
    size_t first_index_to_be_lowered = source8.size();
    for (size_t i = 0; i < source8.size(); ++i) {
      const LChar ch = source8[i];
      if (IsASCIIUpper(ch) || ch & ~0x7F) [[unlikely]] {
        first_index_to_be_lowered = i;
        break;
      }
    }

    // Nothing to do if the string is all ASCII with no uppercase.
    if (first_index_to_be_lowered == source8.size()) {
      return source;
    }

    base::span<LChar> data8;
    scoped_refptr<StringImpl> new_impl =
        StringImpl::CreateUninitialized(source8.size(), data8);

    auto [source8_already_lowercase, source8_tail] =
        source8.split_at(first_index_to_be_lowered);
    auto [data8_already_lowercase, data8_tail] =
        data8.split_at(first_index_to_be_lowered);

    data8_already_lowercase.copy_from(source8_already_lowercase);

    for (size_t i = 0; i < source8_tail.size(); ++i) {
      const LChar ch = source8_tail[i];
      LChar lowered_ch;
      if (ch & ~0x7F) [[unlikely]] {
        lowered_ch = static_cast<LChar>(unicode::ToLower(ch));
      } else {
        lowered_ch = ToASCIILower(ch);
      }
      data8_tail[i] = lowered_ch;
    }
    return new_impl;
  }

  bool no_upper = true;
  UChar ored = 0;

  const base::span<const UChar> source16 = source->Span16();
  for (size_t i = 0; i < source16.size(); ++i) {
    const UChar ch = source16[i];
    if (IsASCIIUpper(ch)) [[unlikely]] {
      no_upper = false;
    }
    ored |= ch;
  }
  // Nothing to do if the string is all ASCII with no uppercase.
  if (no_upper && !(ored & ~0x7F))
    return source;

  CHECK_LE(source16.size(),
           static_cast<wtf_size_t>(std::numeric_limits<int32_t>::max()));

  if (!(ored & ~0x7F)) {
    base::span<UChar> data16;
    scoped_refptr<StringImpl> new_impl =
        StringImpl::CreateUninitialized(source16.size(), data16);

    for (size_t i = 0; i < source16.size(); ++i) {
      data16[i] = ToASCIILower(source16[i]);
    }
    return new_impl;
  }

  // The fast code path was not able to handle this case.
  return nullptr;
}

scoped_refptr<StringImpl> CaseMap::FastToLowerInvariant(StringImpl* source) {
  // Note: This is a hot function in the Dromaeo benchmark.
  DCHECK(source);
  if (scoped_refptr<StringImpl> result = TryFastToLowerInvariant(source))
    return result;
  const char* locale = "";  // "" = root locale.
  return CaseConvert(CaseMapType::kLower, source, locale);
}

scoped_refptr<StringImpl> CaseMap::ToLowerInvariant(StringImpl* source,
                                                    TextOffsetMap* offset_map) {
  DCHECK(source);
  DCHECK(!offset_map || offset_map->IsEmpty());
  if (scoped_refptr<StringImpl> result = TryFastToLowerInvariant(source))
    return result;
  const char* locale = "";  // "" = root locale.
  return CaseConvert(CaseMapType::kLower, source, locale, offset_map);
}

scoped_refptr<StringImpl> CaseMap::ToUpperInvariant(StringImpl* source,
                                                    TextOffsetMap* offset_map) {
  DCHECK(source);
  DCHECK(!offset_map || offset_map->IsEmpty());

  // This function could be optimized for no-op cases the way LowerUnicode() is,
  // but in empirical testing, few actual calls to UpperUnicode() are no-ops, so
  // it wouldn't be worth the extra time for pre-scanning.

  CHECK_LE(source->length(),
           static_cast<wtf_size_t>(std::numeric_limits<int32_t>::max()));

  if (source->Is8Bit()) {
    const base::span<const LChar> source8 = source->Span8();
    base::span<LChar> data8;
    scoped_refptr<StringImpl> new_impl =
        StringImpl::CreateUninitialized(source8.size(), data8);

    // Do a faster loop for the case where all the characters are ASCII.
    LChar ored = 0;
    for (size_t i = 0; i < source8.size(); ++i) {
      const LChar c = source8[i];
      ored |= c;
      data8[i] = ToASCIIUpper(c);
    }
    if (!(ored & ~0x7F))
      return new_impl;

    // Do a slower implementation for cases that include non-ASCII Latin-1
    // characters.
    size_t count_sharp_s_characters = 0;

    // There are two special cases.
    //  1. latin-1 characters when converted to upper case are 16 bit
    //     characters.
    //  2. Lower case sharp-S converts to "SS" (two characters)
    for (size_t i = 0; i < source8.size(); ++i) {
      const LChar c = source8[i];
      if (c == kSmallLetterSharpSCharacter) [[unlikely]] {
        ++count_sharp_s_characters;
      }
      const UChar upper = static_cast<UChar>(unicode::ToUpper(c));
      if (upper > 0xff) [[unlikely]] {
        // Since this upper-cased character does not fit in an 8-bit string, we
        // need to take the 16-bit path.
        goto upconvert;
      }
      data8[i] = static_cast<LChar>(upper);
    }

    if (!count_sharp_s_characters) {
      return new_impl;
    }

    // We have numberSSCharacters sharp-s characters, but none of the other
    // special characters.
    new_impl = StringImpl::CreateUninitialized(
        source8.size() + count_sharp_s_characters, data8);

    size_t dest_index = 0;
    for (size_t i = 0; i < source8.size(); ++i) {
      const LChar c = source8[i];
      if (c == kSmallLetterSharpSCharacter) {
        data8[dest_index++] = 'S';
        data8[dest_index++] = 'S';
        if (offset_map)
          offset_map->Append(i + 1, dest_index);
      } else {
        data8[dest_index++] = static_cast<LChar>(unicode::ToUpper(c));
      }
    }
    return new_impl;
  }

upconvert:
  scoped_refptr<StringImpl> upconverted = source->UpconvertedString();
  base::span<const UChar> source16 = upconverted->Span16();

  base::span<UChar> data16;
  scoped_refptr<StringImpl> new_impl =
      StringImpl::CreateUninitialized(source16.size(), data16);

  // Do a faster loop for the case where all the characters are ASCII.
  UChar ored = 0;
  for (size_t i = 0; i < source16.size(); ++i) {
    const UChar c = source16[i];
    ored |= c;
    data16[i] = ToASCIIUpper(c);
  }
  if (!(ored & ~0x7F))
    return new_impl;

  // Do a slower implementation for cases that include non-ASCII characters.
  const char* locale = "";  // "" = root locale.
  return CaseConvert(CaseMapType::kUpper, source, locale, offset_map);
}

scoped_refptr<StringImpl> CaseMap::ToLower(StringImpl* source,
                                           TextOffsetMap* offset_map) const {
  DCHECK(source);
  DCHECK(!offset_map || offset_map->IsEmpty());

  if (!case_map_locale_)
    return ToLowerInvariant(source, offset_map);
  return CaseConvert(CaseMapType::kLower, source, case_map_locale_, offset_map);
}

scoped_refptr<StringImpl> CaseMap::ToUpper(StringImpl* source,
                                           TextOffsetMap* offset_map) const {
  DCHECK(source);
  DCHECK(!offset_map || offset_map->IsEmpty());

  if (!case_map_locale_)
    return ToUpperInvariant(source, offset_map);
  return CaseConvert(CaseMapType::kUpper, source, case_map_locale_, offset_map);
}

String CaseMap::ToLower(const String& source, TextOffsetMap* offset_map) const {
  DCHECK(!offset_map || offset_map->IsEmpty());

  if (StringImpl* impl = source.Impl())
    return ToLower(impl, offset_map);
  return String();
}

String CaseMap::ToUpper(const String& source, TextOffsetMap* offset_map) const {
  DCHECK(!offset_map || offset_map->IsEmpty());

  if (StringImpl* impl = source.Impl())
    return ToUpper(impl, offset_map);
  return String();
}

String CaseMap::ToTitle(const String& source,
                        TextOffsetMap* offset_map,
                        UChar previous_character) const {
  DCHECK(!offset_map || offset_map->IsEmpty());

  if (StringImpl* impl = source.Impl()) {
    return CaseConvert(CaseMapType::kTitle, impl, case_map_locale_, offset_map,
                       previous_character);
  }
  return String();
}

}  // namespace WTF