summaryrefslogtreecommitdiff
path: root/dom/base/nsTextFragment.cpp
blob: 0ea69bba468133f6bfdc4328ca5aca6864b1411d (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
/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, You can obtain one at http://mozilla.org/MPL/2.0/. */

/*
 * A class which represents a fragment of text (eg inside a text
 * node); if only codepoints below 256 are used, the text is stored as
 * a char*; otherwise the text is stored as a char16_t*
 */

#include "nsTextFragment.h"
#include "nsCRT.h"
#include "nsReadableUtils.h"
#include "nsMemory.h"
#include "nsBidiUtils.h"
#include "nsUnicharUtils.h"
#include "nsUTF8Utils.h"
#include "mozilla/CheckedInt.h"
#include "mozilla/MemoryReporting.h"
#include "mozilla/SSE.h"
#include "nsTextFragmentImpl.h"
#include <algorithm>

#define TEXTFRAG_WHITE_AFTER_NEWLINE 50
#define TEXTFRAG_MAX_NEWLINES 7

// Static buffer used for common fragments
static char* sSpaceSharedString[TEXTFRAG_MAX_NEWLINES + 1];
static char* sTabSharedString[TEXTFRAG_MAX_NEWLINES + 1];
static char sSingleCharSharedString[256];

using mozilla::CheckedUint32;

// static
nsresult
nsTextFragment::Init()
{
  // Create whitespace strings
  uint32_t i;
  for (i = 0; i <= TEXTFRAG_MAX_NEWLINES; ++i) {
    sSpaceSharedString[i] = new char[1 + i + TEXTFRAG_WHITE_AFTER_NEWLINE];
    sTabSharedString[i] = new char[1 + i + TEXTFRAG_WHITE_AFTER_NEWLINE];
    sSpaceSharedString[i][0] = ' ';
    sTabSharedString[i][0] = ' ';
    uint32_t j;
    for (j = 1; j < 1 + i; ++j) {
      sSpaceSharedString[i][j] = '\n';
      sTabSharedString[i][j] = '\n';
    }
    for (; j < (1 + i + TEXTFRAG_WHITE_AFTER_NEWLINE); ++j) {
      sSpaceSharedString[i][j] = ' ';
      sTabSharedString[i][j] = '\t';
    }
  }

  // Create single-char strings
  for (i = 0; i < 256; ++i) {
    sSingleCharSharedString[i] = i;
  }

  return NS_OK;
}

// static
void
nsTextFragment::Shutdown()
{
  uint32_t  i;
  for (i = 0; i <= TEXTFRAG_MAX_NEWLINES; ++i) {
    delete [] sSpaceSharedString[i];
    delete [] sTabSharedString[i];
    sSpaceSharedString[i] = nullptr;
    sTabSharedString[i] = nullptr;
  }
}

nsTextFragment::~nsTextFragment()
{
  ReleaseText();
  MOZ_COUNT_DTOR(nsTextFragment);
}

void
nsTextFragment::ReleaseText()
{
  if (mState.mLength && m1b && mState.mInHeap) {
    free(m2b); // m1b == m2b as far as free is concerned
  }

  m1b = nullptr;
  mState.mIsBidi = false;

  // Set mState.mIs2b, mState.mInHeap, and mState.mLength = 0 with mAllBits;
  mAllBits = 0;
}

nsTextFragment&
nsTextFragment::operator=(const nsTextFragment& aOther)
{
  ReleaseText();

  if (aOther.mState.mLength) {
    if (!aOther.mState.mInHeap) {
      m1b = aOther.m1b; // This will work even if aOther is using m2b
    }
    else {
      CheckedUint32 m2bSize = aOther.mState.mLength;
      m2bSize *= (aOther.mState.mIs2b ? sizeof(char16_t) : sizeof(char));
      m2b = nullptr;
      if (m2bSize.isValid()) {
        m2b = static_cast<char16_t*>(malloc(m2bSize.value()));
      }

      if (m2b) {
        memcpy(m2b, aOther.m2b, m2bSize.value());
      } else {
        // allocate a buffer for a single REPLACEMENT CHARACTER
        m2b = static_cast<char16_t*>(moz_xmalloc(sizeof(char16_t)));
        m2b[0] = 0xFFFD; // REPLACEMENT CHARACTER
        mState.mIs2b = true;
        mState.mInHeap = true;
        mState.mLength = 1;
      }
    }

    if (m1b) {
      mAllBits = aOther.mAllBits;
    }
  }

  return *this;
}

static inline int32_t
FirstNon8BitUnvectorized(const char16_t *str, const char16_t *end)
{
  typedef Non8BitParameters<sizeof(size_t)> p;
  const size_t mask = p::mask();
  const uint32_t alignMask = p::alignMask();
  const uint32_t numUnicharsPerWord = p::numUnicharsPerWord();
  const int32_t len = end - str;
  int32_t i = 0;

  // Align ourselves to a word boundary.
  int32_t alignLen =
    std::min(len, int32_t(((-NS_PTR_TO_INT32(str)) & alignMask) / sizeof(char16_t)));
  for (; i < alignLen; i++) {
    if (str[i] > 255)
      return i;
  }

  // Check one word at a time.
  const int32_t wordWalkEnd = ((len - i) / numUnicharsPerWord) * numUnicharsPerWord;
  for (; i < wordWalkEnd; i += numUnicharsPerWord) {
    const size_t word = *reinterpret_cast<const size_t*>(str + i);
    if (word & mask)
      return i;
  }

  // Take care of the remainder one character at a time.
  for (; i < len; i++) {
    if (str[i] > 255)
      return i;
  }

  return -1;
}

#ifdef MOZILLA_MAY_SUPPORT_SSE2
namespace mozilla {
  namespace SSE2 {
    int32_t FirstNon8Bit(const char16_t *str, const char16_t *end);
  } // namespace SSE2
} // namespace mozilla
#endif

/*
 * This function returns -1 if all characters in str are 8 bit characters.
 * Otherwise, it returns a value less than or equal to the index of the first
 * non-8bit character in str. For example, if first non-8bit character is at
 * position 25, it may return 25, or for example 24, or 16. But it guarantees
 * there is no non-8bit character before returned value.
 */
static inline int32_t
FirstNon8Bit(const char16_t *str, const char16_t *end)
{
#ifdef MOZILLA_MAY_SUPPORT_SSE2
  if (mozilla::supports_sse2()) {
    return mozilla::SSE2::FirstNon8Bit(str, end);
  }
#endif

  return FirstNon8BitUnvectorized(str, end);
}

bool
nsTextFragment::SetTo(const char16_t* aBuffer, int32_t aLength, bool aUpdateBidi)
{
  if (MOZ_UNLIKELY(aLength < 0 || static_cast<uint32_t>(aLength) >
                                      NS_MAX_TEXT_FRAGMENT_LENGTH)) {
    return false;
  }

  ReleaseText();

  if (aLength == 0) {
    return true;
  }
  
  char16_t firstChar = *aBuffer;
  if (aLength == 1 && firstChar < 256) {
    m1b = sSingleCharSharedString + firstChar;
    mState.mInHeap = false;
    mState.mIs2b = false;
    mState.mLength = 1;

    return true;
  }

  const char16_t *ucp = aBuffer;
  const char16_t *uend = aBuffer + aLength;

  // Check if we can use a shared string
  if (aLength <= 1 + TEXTFRAG_WHITE_AFTER_NEWLINE + TEXTFRAG_MAX_NEWLINES &&
     (firstChar == ' ' || firstChar == '\n' || firstChar == '\t')) {
    if (firstChar == ' ') {
      ++ucp;
    }

    const char16_t* start = ucp;
    while (ucp < uend && *ucp == '\n') {
      ++ucp;
    }
    const char16_t* endNewLine = ucp;

    char16_t space = ucp < uend && *ucp == '\t' ? '\t' : ' ';
    while (ucp < uend && *ucp == space) {
      ++ucp;
    }

    if (ucp == uend &&
        endNewLine - start <= TEXTFRAG_MAX_NEWLINES &&
        ucp - endNewLine <= TEXTFRAG_WHITE_AFTER_NEWLINE) {
      char** strings = space == ' ' ? sSpaceSharedString : sTabSharedString;
      m1b = strings[endNewLine - start];

      // If we didn't find a space in the beginning, skip it now.
      if (firstChar != ' ') {
        ++m1b;
      }

      mState.mInHeap = false;
      mState.mIs2b = false;
      mState.mLength = aLength;

      return true;        
    }
  }

  // See if we need to store the data in ucs2 or not
  int32_t first16bit = FirstNon8Bit(ucp, uend);

  if (first16bit != -1) { // aBuffer contains no non-8bit character
    // Use ucs2 storage because we have to
    CheckedUint32 m2bSize = aLength;
    m2bSize *= sizeof(char16_t);
    if (!m2bSize.isValid()) {
      return false;
    }

    m2b = static_cast<char16_t*>(malloc(m2bSize.value()));
    if (!m2b) {
      return false;
    }
    memcpy(m2b, aBuffer, m2bSize.value());

    mState.mIs2b = true;
    if (aUpdateBidi) {
      UpdateBidiFlag(aBuffer + first16bit, aLength - first16bit);
    }

  } else {
    // Use 1 byte storage because we can
    char* buff = static_cast<char*>(malloc(aLength));
    if (!buff) {
      return false;
    }

    // Copy data
    LossyConvertEncoding16to8 converter(buff);
    copy_string(aBuffer, aBuffer+aLength, converter);
    m1b = buff;
    mState.mIs2b = false;
  }

  // Setup our fields
  mState.mInHeap = true;
  mState.mLength = aLength;

  return true;
}

void
nsTextFragment::CopyTo(char16_t *aDest, int32_t aOffset, int32_t aCount)
{
  NS_ASSERTION(aOffset >= 0, "Bad offset passed to nsTextFragment::CopyTo()!");
  NS_ASSERTION(aCount >= 0, "Bad count passed to nsTextFragment::CopyTo()!");

  if (aOffset < 0) {
    aOffset = 0;
  }

  if (uint32_t(aOffset + aCount) > GetLength()) {
    aCount = mState.mLength - aOffset;
  }

  if (aCount != 0) {
    if (mState.mIs2b) {
      memcpy(aDest, m2b + aOffset, sizeof(char16_t) * aCount);
    } else {
      const char *cp = m1b + aOffset;
      const char *end = cp + aCount;
      LossyConvertEncoding8to16 converter(aDest);
      copy_string(cp, end, converter);
    }
  }
}

bool
nsTextFragment::Append(const char16_t* aBuffer, uint32_t aLength, bool aUpdateBidi)
{
  // This is a common case because some callsites create a textnode
  // with a value by creating the node and then calling AppendData.
  if (mState.mLength == 0) {
    if (MOZ_UNLIKELY(aLength > INT32_MAX)) {
      return false;
    }
    return SetTo(aBuffer, aLength, aUpdateBidi);
  }

  // Should we optimize for aData.Length() == 0?

  // Note: Using CheckedInt here is wrong as nsTextFragment is 29 bits and needs an
  // explicit check for that length and not INT_MAX. Also, this method can be a very
  // hot path and cause performance loss since CheckedInt isn't inlined.
  if (NS_MAX_TEXT_FRAGMENT_LENGTH - mState.mLength < aLength) {
    return false;  // Would be overflowing if we'd continue.
  }

  if (mState.mIs2b) {
    size_t size = mState.mLength + aLength;
    if (SIZE_MAX / sizeof(char16_t) < size) {
      return false;  // Would be overflowing if we'd continue.
    }
    size *= sizeof(char16_t);

    // Already a 2-byte string so the result will be too
    char16_t* buff = static_cast<char16_t*>(realloc(m2b, size));
    if (!buff) {
      return false;
    }

    memcpy(buff + mState.mLength, aBuffer, aLength * sizeof(char16_t));
    mState.mLength += aLength;
    m2b = buff;

    if (aUpdateBidi) {
      UpdateBidiFlag(aBuffer, aLength);
    }

    return true;
  }

  // Current string is a 1-byte string, check if the new data fits in one byte too.
  int32_t first16bit = FirstNon8Bit(aBuffer, aBuffer + aLength);

  if (first16bit != -1) { // aBuffer contains no non-8bit character
    size_t size = mState.mLength + aLength;
    if (SIZE_MAX / sizeof(char16_t) < size) {
      return false;  // Would be overflowing if we'd continue.
    }
    size *= sizeof(char16_t);

    // The old data was 1-byte, but the new is not so we have to expand it
    // all to 2-byte
    char16_t* buff = static_cast<char16_t*>(malloc(size));
    if (!buff) {
      return false;
    }

    // Copy data into buff
    LossyConvertEncoding8to16 converter(buff);
    copy_string(m1b, m1b+mState.mLength, converter);

    memcpy(buff + mState.mLength, aBuffer, aLength * sizeof(char16_t));
    mState.mLength += aLength;
    mState.mIs2b = true;

    if (mState.mInHeap) {
      free(m2b);
    }
    m2b = buff;

    mState.mInHeap = true;

    if (aUpdateBidi) {
      UpdateBidiFlag(aBuffer + first16bit, aLength - first16bit);
    }

    return true;
  }

  // The new and the old data is all 1-byte

  // Note: Using CheckedInt here is wrong. See above.
  if (NS_MAX_TEXT_FRAGMENT_LENGTH - mState.mLength < aLength) {
    return false;  // Would be overflowing if we'd continue.
  }
  
  size_t size = mState.mLength + aLength;
  MOZ_ASSERT(sizeof(char) == 1);
  char* buff;
  if (mState.mInHeap) {
    buff = static_cast<char*>(realloc(const_cast<char*>(m1b), size));
    if (!buff) {
      return false;
    }
  } else {
    buff = static_cast<char*>(malloc(size));
    if (!buff) {
      return false;
    }

    memcpy(buff, m1b, mState.mLength);
    mState.mInHeap = true;
  }

  // Copy aBuffer into buff.
  LossyConvertEncoding16to8 converter(buff + mState.mLength);
  copy_string(aBuffer, aBuffer + aLength, converter);

  m1b = buff;
  mState.mLength += aLength;

  return true;
}

/* virtual */ size_t
nsTextFragment::SizeOfExcludingThis(mozilla::MallocSizeOf aMallocSizeOf) const
{
  if (Is2b()) {
    return aMallocSizeOf(m2b);
  }

  if (mState.mInHeap) {
    return aMallocSizeOf(m1b);
  }

  return 0;
}

// To save time we only do this when we really want to know, not during
// every allocation
void
nsTextFragment::UpdateBidiFlag(const char16_t* aBuffer, uint32_t aLength)
{
  if (mState.mIs2b && !mState.mIsBidi) {
    const char16_t* cp = aBuffer;
    const char16_t* end = cp + aLength;
    while (cp < end) {
      char16_t ch1 = *cp++;
      uint32_t utf32Char = ch1;
      if (NS_IS_HIGH_SURROGATE(ch1) &&
          cp < end &&
          NS_IS_LOW_SURROGATE(*cp)) {
        char16_t ch2 = *cp++;
        utf32Char = SURROGATE_TO_UCS4(ch1, ch2);
      }
      if (UTF32_CHAR_IS_BIDI(utf32Char) || IsBidiControlRTL(utf32Char)) {
        mState.mIsBidi = true;
        break;
      }
    }
  }
}