third_party/WebKit/Source/wtf/text/WTFString.h

   1 /*
   2  * (C) 1999 Lars Knoll (knoll@kde.org)
   3  * Copyright (C) 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2012, 2013 Apple Inc. All rights reserved.
   4  *
   5  * This library is free software; you can redistribute it and/or
   6  * modify it under the terms of the GNU Library General Public
   7  * License as published by the Free Software Foundation; either
   8  * version 2 of the License, or (at your option) any later version.
   9  *
  10  * This library is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13  * Library General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU Library General Public License
  16  * along with this library; see the file COPYING.LIB.  If not, write to
  17  * the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
  18  * Boston, MA 02110-1301, USA.
  19  *
  20  */
  21
  22 #ifndef WTFString_h
  23 #define WTFString_h
  24
  25 // This file would be called String.h, but that conflicts with <string.h>
  26 // on systems without case-sensitive file systems.
  27
  28 #include "wtf/HashTableDeletedValueType.h"
  29 #include "wtf/WTFExport.h"
  30 #include "wtf/testing/WTFUnitTestHelpersExport.h"
  31 #include "wtf/text/ASCIIFastPath.h"
  32 #include "wtf/text/StringImpl.h"
  33 #include "wtf/text/StringView.h"
  34 #include <algorithm>
  35 #include <iosfwd>
  36
  37 #ifdef __OBJC__
  38 #include <objc/objc.h>
  39 #endif
  40
  41 namespace WTF {
  42
  43 class CString;
  44 struct StringHash;
  45
  46 // Declarations of string operations
  47
  48 WTF_EXPORT int charactersToIntStrict(const LChar*, size_t, bool* ok = 0, int base = 10);
  49 WTF_EXPORT int charactersToIntStrict(const UChar*, size_t, bool* ok = 0, int base = 10);
  50 WTF_EXPORT unsigned charactersToUIntStrict(const LChar*, size_t, bool* ok = 0, int base = 10);
  51 WTF_EXPORT unsigned charactersToUIntStrict(const UChar*, size_t, bool* ok = 0, int base = 10);
  52 WTF_EXPORT int64_t charactersToInt64Strict(const LChar*, size_t, bool* ok = 0, int base = 10);
  53 WTF_EXPORT int64_t charactersToInt64Strict(const UChar*, size_t, bool* ok = 0, int base = 10);
  54 WTF_EXPORT uint64_t charactersToUInt64Strict(const LChar*, size_t, bool* ok = 0, int base = 10);
  55 WTF_EXPORT uint64_t charactersToUInt64Strict(const UChar*, size_t, bool* ok = 0, int base = 10);
  56
  57 WTF_EXPORT int charactersToInt(const LChar*, size_t, bool* ok = 0); // ignores trailing garbage
  58 WTF_EXPORT int charactersToInt(const UChar*, size_t, bool* ok = 0); // ignores trailing garbage
  59 WTF_EXPORT unsigned charactersToUInt(const LChar*, size_t, bool* ok = 0); // ignores trailing garbage
  60 WTF_EXPORT unsigned charactersToUInt(const UChar*, size_t, bool* ok = 0); // ignores trailing garbage
  61 WTF_EXPORT int64_t charactersToInt64(const LChar*, size_t, bool* ok = 0); // ignores trailing garbage
  62 WTF_EXPORT int64_t charactersToInt64(const UChar*, size_t, bool* ok = 0); // ignores trailing garbage
  63 WTF_EXPORT uint64_t charactersToUInt64(const LChar*, size_t, bool* ok = 0); // ignores trailing garbage
  64 WTF_EXPORT uint64_t charactersToUInt64(const UChar*, size_t, bool* ok = 0); // ignores trailing garbage
  65
  66 // FIXME: Like the strict functions above, these give false for "ok" when there is trailing garbage.
  67 // Like the non-strict functions above, these return the value when there is trailing garbage.
  68 // It would be better if these were more consistent with the above functions instead.
  69 WTF_EXPORT double charactersToDouble(const LChar*, size_t, bool* ok = 0);
  70 WTF_EXPORT double charactersToDouble(const UChar*, size_t, bool* ok = 0);
  71 WTF_EXPORT float charactersToFloat(const LChar*, size_t, bool* ok = 0);
  72 WTF_EXPORT float charactersToFloat(const UChar*, size_t, bool* ok = 0);
  73 WTF_EXPORT float charactersToFloat(const LChar*, size_t, size_t& parsedLength);
  74 WTF_EXPORT float charactersToFloat(const UChar*, size_t, size_t& parsedLength);
  75
  76 enum TrailingZerosTruncatingPolicy {
  77     KeepTrailingZeros,
  78     TruncateTrailingZeros
  79 };
  80
  81 enum UTF8ConversionMode {
  82     LenientUTF8Conversion,
  83     StrictUTF8Conversion,
  84     StrictUTF8ConversionReplacingUnpairedSurrogatesWithFFFD
  85 };
  86
  87 template<bool isSpecialCharacter(UChar), typename CharacterType>
  88 bool isAllSpecialCharacters(const CharacterType*, size_t);
  89
  90 // You can find documentation about this class in this doc:
  91 // https://docs.google.com/document/d/1kOCUlJdh2WJMJGDf-WoEQhmnjKLaOYRbiHz5TiGJl14/edit?usp=sharing
  92 class WTF_EXPORT String {
  93 public:
  94     // Construct a null string, distinguishable from an empty string.
  95     String() { }
  96
  97     // Construct a string with UTF-16 data.
  98     String(const UChar* characters, unsigned length);
  99
 100     // Construct a string by copying the contents of a vector.
 101     // This method will never create a null string. Vectors with size() == 0
 102     // will return the empty string.
 103     // NOTE: This is different from String(vector.data(), vector.size())
 104     // which will sometimes return a null string when vector.data() is null
 105     // which can only occur for vectors without inline capacity.
 106     // See: https://bugs.webkit.org/show_bug.cgi?id=109792
 107     template<size_t inlineCapacity>
 108     explicit String(const Vector<UChar, inlineCapacity>&);
 109
 110     // Construct a string with UTF-16 data, from a null-terminated source.
 111     String(const UChar*);
 112
 113     // Construct a string with latin1 data.
 114     String(const LChar* characters, unsigned length);
 115     String(const char* characters, unsigned length);
 116
 117     // Construct a string with latin1 data, from a null-terminated source.
 118     String(const LChar* characters);
 119     String(const char* characters);
 120
 121     // Construct a string referencing an existing StringImpl.
 122     String(StringImpl* impl) : m_impl(impl) { }
 123     String(PassRefPtr<StringImpl> impl) : m_impl(impl) { }
 124
 125     void swap(String& o) { m_impl.swap(o.m_impl); }
 126
 127     template<typename CharType>
 128     static String adopt(StringBuffer<CharType>& buffer)
 129     {
 130         if (!buffer.length())
 131             return StringImpl::empty();
 132         return String(buffer.release());
 133     }
 134
 135     bool isNull() const { return !m_impl; }
 136     bool isEmpty() const { return !m_impl || !m_impl->length(); }
 137
 138     StringImpl* impl() const { return m_impl.get(); }
 139     PassRefPtr<StringImpl> releaseImpl() { return m_impl.release(); }
 140
 141     unsigned length() const
 142     {
 143         if (!m_impl)
 144             return 0;
 145         return m_impl->length();
 146     }
 147
 148     const LChar* characters8() const
 149     {
 150         if (!m_impl)
 151             return 0;
 152         ASSERT(m_impl->is8Bit());
 153         return m_impl->characters8();
 154     }
 155
 156     const UChar* characters16() const
 157     {
 158         if (!m_impl)
 159             return 0;
 160         ASSERT(!m_impl->is8Bit());
 161         return m_impl->characters16();
 162     }
 163
 164     // Return characters8() or characters16() depending on CharacterType.
 165     template <typename CharacterType>
 166     inline const CharacterType* getCharacters() const;
 167
 168     bool is8Bit() const { return m_impl->is8Bit(); }
 169
 170     unsigned sizeInBytes() const
 171     {
 172         if (!m_impl)
 173             return 0;
 174         return m_impl->length() * (is8Bit() ? sizeof(LChar) : sizeof(UChar));
 175     }
 176
 177     CString ascii() const;
 178     CString latin1() const;
 179     CString utf8(UTF8ConversionMode = LenientUTF8Conversion) const;
 180
 181     UChar operator[](unsigned index) const
 182     {
 183         if (!m_impl || index >= m_impl->length())
 184             return 0;
 185         return (*m_impl)[index];
 186     }
 187
 188     static String number(int);
 189     static String number(unsigned);
 190     static String number(long);
 191     static String number(unsigned long);
 192     static String number(long long);
 193     static String number(unsigned long long);
 194
 195     static String number(double, unsigned precision = 6, TrailingZerosTruncatingPolicy = TruncateTrailingZeros);
 196
 197     // Number to String conversion following the ECMAScript definition.
 198     static String numberToStringECMAScript(double);
 199     static String numberToStringFixedWidth(double, unsigned decimalPlaces);
 200
 201     // Find a single character or string, also with match function & latin1 forms.
 202     size_t find(UChar c, unsigned start = 0) const
 203         { return m_impl ? m_impl->find(c, start) : kNotFound; }
 204     size_t find(LChar c, unsigned start = 0) const
 205         { return m_impl ? m_impl->find(c, start) : kNotFound; }
 206     size_t find(char c, unsigned start = 0) const { return find(static_cast<LChar>(c), start); }
 207
 208     size_t find(const String& str) const
 209         { return m_impl ? m_impl->find(str.impl()) : kNotFound; }
 210     size_t find(const String& str, unsigned start) const
 211         { return m_impl ? m_impl->find(str.impl(), start) : kNotFound; }
 212
 213     size_t find(CharacterMatchFunctionPtr matchFunction, unsigned start = 0) const
 214         { return m_impl ? m_impl->find(matchFunction, start) : kNotFound; }
 215     size_t find(const LChar* str, unsigned start = 0) const
 216         { return m_impl ? m_impl->find(str, start) : kNotFound; }
 217
 218     size_t findNextLineStart(unsigned start = 0) const
 219         { return m_impl ? m_impl->findNextLineStart(start) : kNotFound; }
 220
 221     // Find the last instance of a single character or string.
 222     size_t reverseFind(UChar c, unsigned start = UINT_MAX) const
 223         { return m_impl ? m_impl->reverseFind(c, start) : kNotFound; }
 224     size_t reverseFind(const String& str, unsigned start = UINT_MAX) const
 225         { return m_impl ? m_impl->reverseFind(str.impl(), start) : kNotFound; }
 226
 227     // Case insensitive string matching.
 228     size_t findIgnoringCase(const LChar* str, unsigned start = 0) const
 229         { return m_impl ? m_impl->findIgnoringCase(str, start) : kNotFound; }
 230     size_t findIgnoringCase(const String& str, unsigned start = 0) const
 231         { return m_impl ? m_impl->findIgnoringCase(str.impl(), start) : kNotFound; }
 232     size_t reverseFindIgnoringCase(const String& str, unsigned start = UINT_MAX) const
 233         { return m_impl ? m_impl->reverseFindIgnoringCase(str.impl(), start) : kNotFound; }
 234
 235     // Wrappers for find & reverseFind adding dynamic sensitivity check.
 236     size_t find(const LChar* str, unsigned start, TextCaseSensitivity caseSensitivity) const
 237         { return (caseSensitivity == TextCaseSensitive) ? find(str, start) : findIgnoringCase(str, start); }
 238     size_t find(const String& str, unsigned start, TextCaseSensitivity caseSensitivity) const
 239         { return (caseSensitivity == TextCaseSensitive) ? find(str, start) : findIgnoringCase(str, start); }
 240     size_t reverseFind(const String& str, unsigned start, TextCaseSensitivity caseSensitivity) const
 241         { return (caseSensitivity == TextCaseSensitive) ? reverseFind(str, start) : reverseFindIgnoringCase(str, start); }
 242
 243     Vector<UChar> charactersWithNullTermination() const;
 244     unsigned copyTo(UChar* buffer, unsigned pos, unsigned maxLength) const;
 245
 246     template<size_t inlineCapacity>
 247     void appendTo(Vector<UChar, inlineCapacity>&, unsigned pos = 0, unsigned len = UINT_MAX) const;
 248
 249     template<typename BufferType>
 250     void appendTo(BufferType&, unsigned pos = 0, unsigned len = UINT_MAX) const;
 251
 252     template<size_t inlineCapacity>
 253     void prependTo(Vector<UChar, inlineCapacity>&, unsigned pos = 0, unsigned len = UINT_MAX) const;
 254
 255     UChar32 characterStartingAt(unsigned) const;
 256     template<typename CharacterType>
 257     bool contains(CharacterType c) const { return find(c) != kNotFound; }
 258     bool contains(const LChar* str, TextCaseSensitivity caseSensitivity = TextCaseSensitive) const { return find(str, 0, caseSensitivity) != kNotFound; }
 259     bool contains(const String& str, TextCaseSensitivity caseSensitivity = TextCaseSensitive) const { return find(str, 0, caseSensitivity) != kNotFound; }
 260
 261     bool startsWith(const String& s, TextCaseSensitivity caseSensitivity = TextCaseSensitive) const
 262         { return m_impl ? m_impl->startsWith(s.impl(), caseSensitivity) : s.isEmpty(); }
 263     bool startsWith(UChar character) const
 264         { return m_impl ? m_impl->startsWith(character) : false; }
 265     template<unsigned matchLength>
 266     bool startsWith(const char (&prefix)[matchLength], TextCaseSensitivity caseSensitivity = TextCaseSensitive) const
 267         { return m_impl ? m_impl->startsWith<matchLength>(prefix, caseSensitivity) : !matchLength; }
 268
 269     bool endsWith(const String& s, TextCaseSensitivity caseSensitivity = TextCaseSensitive) const
 270         { return m_impl ? m_impl->endsWith(s.impl(), caseSensitivity) : s.isEmpty(); }
 271     bool endsWith(UChar character) const
 272         { return m_impl ? m_impl->endsWith(character) : false; }
 273     template<unsigned matchLength>
 274     bool endsWith(const char (&prefix)[matchLength], TextCaseSensitivity caseSensitivity = TextCaseSensitive) const
 275         { return m_impl ? m_impl->endsWith<matchLength>(prefix, caseSensitivity) : !matchLength; }
 276
 277     void append(const String&);
 278     void append(LChar);
 279     void append(char c) { append(static_cast<LChar>(c)); }
 280     void append(UChar);
 281     void append(const LChar*, unsigned length);
 282     void append(const char* charactersToAppend, unsigned length) { append(reinterpret_cast<const LChar*>(charactersToAppend), length); }
 283     void append(const UChar*, unsigned length);
 284     void insert(const String&, unsigned pos);
 285     void insert(const LChar*, unsigned length, unsigned pos);
 286     void insert(const UChar*, unsigned length, unsigned pos);
 287
 288     String& replace(UChar a, UChar b) { if (m_impl) m_impl = m_impl->replace(a, b); return *this; }
 289     String& replace(UChar a, const String& b) { if (m_impl) m_impl = m_impl->replace(a, b.impl()); return *this; }
 290     String& replace(const String& a, const String& b) { if (m_impl) m_impl = m_impl->replace(a.impl(), b.impl()); return *this; }
 291     String& replace(unsigned index, unsigned len, const String& b) { if (m_impl) m_impl = m_impl->replace(index, len, b.impl()); return *this; }
 292
 293     template<unsigned charactersCount>
 294     ALWAYS_INLINE String& replaceWithLiteral(UChar a, const char (&characters)[charactersCount])
 295     {
 296         if (m_impl)
 297             m_impl = m_impl->replace(a, characters, charactersCount - 1);
 298
 299         return *this;
 300     }
 301
 302     void fill(UChar c) { if (m_impl) m_impl = m_impl->fill(c); }
 303
 304     void ensure16Bit();
 305
 306     void truncate(unsigned len);
 307     void remove(unsigned pos, int len = 1);
 308
 309     String substring(unsigned pos, unsigned len = UINT_MAX) const;
 310     String left(unsigned len) const { return substring(0, len); }
 311     String right(unsigned len) const { return substring(length() - len, len); }
 312
 313     StringView createView() const { return StringView(impl()); }
 314     StringView createView(unsigned offset, unsigned length) const { return StringView(impl(), offset, length); }
 315
 316     // Returns a lowercase/uppercase version of the string
 317     String lower() const;
 318     String upper() const;
 319
 320     String lower(const AtomicString& localeIdentifier) const;
 321     String upper(const AtomicString& localeIdentifier) const;
 322
 323     String stripWhiteSpace() const;
 324     String stripWhiteSpace(IsWhiteSpaceFunctionPtr) const;
 325     String simplifyWhiteSpace(StripBehavior stripBehavior = StripExtraWhiteSpace) const;
 326     String simplifyWhiteSpace(IsWhiteSpaceFunctionPtr, StripBehavior stripBehavior = StripExtraWhiteSpace) const;
 327
 328     String removeCharacters(CharacterMatchFunctionPtr) const;
 329     template<bool isSpecialCharacter(UChar)> bool isAllSpecialCharacters() const;
 330
 331     // Return the string with case folded for case insensitive comparison.
 332     String foldCase() const;
 333
 334     static String format(const char *, ...) WTF_ATTRIBUTE_PRINTF(1, 2);
 335
 336     // Returns an uninitialized string. The characters needs to be written
 337     // into the buffer returned in data before the returned string is used.
 338     // Failure to do this will have unpredictable results.
 339     static String createUninitialized(unsigned length, UChar*& data) { return StringImpl::createUninitialized(length, data); }
 340     static String createUninitialized(unsigned length, LChar*& data) { return StringImpl::createUninitialized(length, data); }
 341
 342     void split(const String& separator, bool allowEmptyEntries, Vector<String>& result) const;
 343     void split(const String& separator, Vector<String>& result) const
 344     {
 345         split(separator, false, result);
 346     }
 347     void split(UChar separator, bool allowEmptyEntries, Vector<String>& result) const;
 348     void split(UChar separator, Vector<String>& result) const
 349     {
 350         split(separator, false, result);
 351     }
 352
 353     int toIntStrict(bool* ok = 0, int base = 10) const;
 354     unsigned toUIntStrict(bool* ok = 0, int base = 10) const;
 355     int64_t toInt64Strict(bool* ok = 0, int base = 10) const;
 356     uint64_t toUInt64Strict(bool* ok = 0, int base = 10) const;
 357
 358     int toInt(bool* ok = 0) const;
 359     unsigned toUInt(bool* ok = 0) const;
 360     int64_t toInt64(bool* ok = 0) const;
 361     uint64_t toUInt64(bool* ok = 0) const;
 362
 363     // FIXME: Like the strict functions above, these give false for "ok" when there is trailing garbage.
 364     // Like the non-strict functions above, these return the value when there is trailing garbage.
 365     // It would be better if these were more consistent with the above functions instead.
 366     double toDouble(bool* ok = 0) const;
 367     float toFloat(bool* ok = 0) const;
 368
 369     String isolatedCopy() const;
 370     bool isSafeToSendToAnotherThread() const;
 371
 372 #if USE(CF)
 373     String(CFStringRef);
 374     RetainPtr<CFStringRef> createCFString() const;
 375 #endif
 376
 377 #ifdef __OBJC__
 378     String(NSString*);
 379
 380     // This conversion maps NULL to "", which loses the meaning of NULL, but we
 381     // need this mapping because AppKit crashes when passed nil NSStrings.
 382     operator NSString*() const { if (!m_impl) return @""; return *m_impl; }
 383 #endif
 384
 385     static String make8BitFrom16BitSource(const UChar*, size_t);
 386     template<size_t inlineCapacity>
 387     static String make8BitFrom16BitSource(const Vector<UChar, inlineCapacity>& buffer)
 388     {
 389         return make8BitFrom16BitSource(buffer.data(), buffer.size());
 390     }
 391
 392     static String make16BitFrom8BitSource(const LChar*, size_t);
 393
 394     // String::fromUTF8 will return a null string if
 395     // the input data contains invalid UTF-8 sequences.
 396     static String fromUTF8(const LChar*, size_t);
 397     static String fromUTF8(const LChar*);
 398     static String fromUTF8(const char* s, size_t length) { return fromUTF8(reinterpret_cast<const LChar*>(s), length); }
 399     static String fromUTF8(const char* s) { return fromUTF8(reinterpret_cast<const LChar*>(s)); }
 400     static String fromUTF8(const CString&);
 401
 402     // Tries to convert the passed in string to UTF-8, but will fall back to Latin-1 if the string is not valid UTF-8.
 403     static String fromUTF8WithLatin1Fallback(const LChar*, size_t);
 404     static String fromUTF8WithLatin1Fallback(const char* s, size_t length) { return fromUTF8WithLatin1Fallback(reinterpret_cast<const LChar*>(s), length); }
 405
 406     bool containsOnlyASCII() const;
 407     bool containsOnlyLatin1() const;
 408     bool containsOnlyWhitespace() const { return !m_impl || m_impl->containsOnlyWhitespace(); }
 409
 410     // Hash table deleted values, which are only constructed and never copied or destroyed.
 411     String(WTF::HashTableDeletedValueType) : m_impl(WTF::HashTableDeletedValue) { }
 412     bool isHashTableDeletedValue() const { return m_impl.isHashTableDeletedValue(); }
 413
 414 #ifndef NDEBUG
 415     void show() const;
 416 #endif
 417
 418     // Workaround for a compiler bug. Use operator[] instead.
 419     UChar characterAt(unsigned index) const
 420     {
 421         if (!m_impl || index >= m_impl->length())
 422             return 0;
 423         return (*m_impl)[index];
 424     }
 425
 426 private:
 427     typedef struct ImplicitConversionFromWTFStringToBoolDisallowed* (String::*UnspecifiedBoolType);
 428     operator UnspecifiedBoolType() const;
 429
 430     template <typename CharacterType>
 431     void removeInternal(const CharacterType*, unsigned, int);
 432
 433     template <typename CharacterType>
 434     void appendInternal(CharacterType);
 435
 436     RefPtr<StringImpl> m_impl;
 437 };
 438
 439 inline bool operator==(const String& a, const String& b) { return equal(a.impl(), b.impl()); }
 440 inline bool operator==(const String& a, const LChar* b) { return equal(a.impl(), b); }
 441 inline bool operator==(const String& a, const char* b) { return equal(a.impl(), reinterpret_cast<const LChar*>(b)); }
 442 inline bool operator==(const LChar* a, const String& b) { return equal(a, b.impl()); }
 443 inline bool operator==(const char* a, const String& b) { return equal(reinterpret_cast<const LChar*>(a), b.impl()); }
 444 template<size_t inlineCapacity>
 445 inline bool operator==(const Vector<char, inlineCapacity>& a, const String& b) { return equal(b.impl(), a.data(), a.size()); }
 446 template<size_t inlineCapacity>
 447 inline bool operator==(const String& a, const Vector<char, inlineCapacity>& b) { return b == a; }
 448
 449
 450 inline bool operator!=(const String& a, const String& b) { return !equal(a.impl(), b.impl()); }
 451 inline bool operator!=(const String& a, const LChar* b) { return !equal(a.impl(), b); }
 452 inline bool operator!=(const String& a, const char* b) { return !equal(a.impl(), reinterpret_cast<const LChar*>(b)); }
 453 inline bool operator!=(const LChar* a, const String& b) { return !equal(a, b.impl()); }
 454 inline bool operator!=(const char* a, const String& b) { return !equal(reinterpret_cast<const LChar*>(a), b.impl()); }
 455 template<size_t inlineCapacity>
 456 inline bool operator!=(const Vector<char, inlineCapacity>& a, const String& b) { return !(a == b); }
 457 template<size_t inlineCapacity>
 458 inline bool operator!=(const String& a, const Vector<char, inlineCapacity>& b) { return b != a; }
 459
 460 inline bool equalIgnoringCase(const String& a, const String& b) { return equalIgnoringCase(a.impl(), b.impl()); }
 461 inline bool equalIgnoringCase(const String& a, const LChar* b) { return equalIgnoringCase(a.impl(), b); }
 462 inline bool equalIgnoringCase(const String& a, const char* b) { return equalIgnoringCase(a.impl(), reinterpret_cast<const LChar*>(b)); }
 463 inline bool equalIgnoringCase(const LChar* a, const String& b) { return equalIgnoringCase(a, b.impl()); }
 464 inline bool equalIgnoringCase(const char* a, const String& b) { return equalIgnoringCase(reinterpret_cast<const LChar*>(a), b.impl()); }
 465
 466 inline bool equalPossiblyIgnoringCase(const String& a, const String& b, bool ignoreCase)
 467 {
 468     return ignoreCase ? equalIgnoringCase(a, b) : (a == b);
 469 }
 470
 471 inline bool equalIgnoringNullity(const String& a, const String& b) { return equalIgnoringNullity(a.impl(), b.impl()); }
 472
 473 template<size_t inlineCapacity>
 474 inline bool equalIgnoringNullity(const Vector<UChar, inlineCapacity>& a, const String& b) { return equalIgnoringNullity(a, b.impl()); }
 475
 476 inline bool operator!(const String& str) { return str.isNull(); }
 477
 478 inline void swap(String& a, String& b) { a.swap(b); }
 479
 480 // Definitions of string operations
 481
 482 template<size_t inlineCapacity>
 483 String::String(const Vector<UChar, inlineCapacity>& vector)
 484     : m_impl(vector.size() ? StringImpl::create(vector.data(), vector.size()) : StringImpl::empty())
 485 {
 486 }
 487
 488 template<>
 489 inline const LChar* String::getCharacters<LChar>() const
 490 {
 491     ASSERT(is8Bit());
 492     return characters8();
 493 }
 494
 495 template<>
 496 inline const UChar* String::getCharacters<UChar>() const
 497 {
 498     ASSERT(!is8Bit());
 499     return characters16();
 500 }
 501
 502 inline bool String::containsOnlyLatin1() const
 503 {
 504     if (isEmpty())
 505         return true;
 506
 507     if (is8Bit())
 508         return true;
 509
 510     const UChar* characters = characters16();
 511     UChar ored = 0;
 512     for (size_t i = 0; i < m_impl->length(); ++i)
 513         ored |= characters[i];
 514     return !(ored & 0xFF00);
 515 }
 516
 517
 518 #ifdef __OBJC__
 519 // This is for situations in WebKit where the long standing behavior has been
 520 // "nil if empty", so we try to maintain longstanding behavior for the sake of
 521 // entrenched clients
 522 inline NSString* nsStringNilIfEmpty(const String& str) {  return str.isEmpty() ? nil : (NSString*)str; }
 523 #endif
 524
 525 inline bool String::containsOnlyASCII() const
 526 {
 527     if (isEmpty())
 528         return true;
 529
 530     if (is8Bit())
 531         return charactersAreAllASCII(characters8(), m_impl->length());
 532
 533     return charactersAreAllASCII(characters16(), m_impl->length());
 534 }
 535
 536 WTF_EXPORT int codePointCompare(const String&, const String&);
 537
 538 inline bool codePointCompareLessThan(const String& a, const String& b)
 539 {
 540     return codePointCompare(a.impl(), b.impl()) < 0;
 541 }
 542
 543 template<size_t inlineCapacity>
 544 inline void append(Vector<UChar, inlineCapacity>& vector, const String& string)
 545 {
 546     unsigned length = string.length();
 547     if (!length)
 548         return;
 549     if (string.is8Bit()) {
 550         const LChar* characters8 = string.characters8();
 551         vector.reserveCapacity(vector.size() + length);
 552         for (size_t i = 0; i < length; ++i)
 553             vector.uncheckedAppend(characters8[i]);
 554     } else {
 555         vector.append(string.characters16(), length);
 556     }
 557 }
 558
 559 template<bool isSpecialCharacter(UChar), typename CharacterType>
 560 inline bool isAllSpecialCharacters(const CharacterType* characters, size_t length)
 561 {
 562     for (size_t i = 0; i < length; ++i) {
 563         if (!isSpecialCharacter(characters[i]))
 564             return false;
 565     }
 566     return true;
 567 }
 568
 569 template<bool isSpecialCharacter(UChar)>
 570 inline bool String::isAllSpecialCharacters() const
 571 {
 572     size_t len = length();
 573
 574     if (!len)
 575         return true;
 576
 577     if (is8Bit())
 578         return WTF::isAllSpecialCharacters<isSpecialCharacter, LChar>(characters8(), len);
 579     return WTF::isAllSpecialCharacters<isSpecialCharacter, UChar>(characters16(), len);
 580 }
 581
 582 template<size_t inlineCapacity>
 583 inline void String::appendTo(Vector<UChar, inlineCapacity>& result, unsigned pos, unsigned len) const
 584 {
 585     unsigned numberOfCharactersToCopy = std::min(len, length() - pos);
 586     if (!numberOfCharactersToCopy)
 587         return;
 588     result.reserveCapacity(result.size() + numberOfCharactersToCopy);
 589     if (is8Bit()) {
 590         const LChar* characters8 = m_impl->characters8();
 591         for (size_t i = 0; i < numberOfCharactersToCopy; ++i)
 592             result.uncheckedAppend(characters8[pos + i]);
 593     } else {
 594         const UChar* characters16 = m_impl->characters16();
 595         result.append(characters16 + pos, numberOfCharactersToCopy);
 596     }
 597 }
 598
 599 template<typename BufferType>
 600 inline void String::appendTo(BufferType& result, unsigned pos, unsigned len) const
 601 {
 602     unsigned numberOfCharactersToCopy = std::min(len, length() - pos);
 603     if (!numberOfCharactersToCopy)
 604         return;
 605     if (is8Bit())
 606         result.append(m_impl->characters8() + pos, numberOfCharactersToCopy);
 607     else
 608         result.append(m_impl->characters16() + pos, numberOfCharactersToCopy);
 609 }
 610
 611 template<size_t inlineCapacity>
 612 inline void String::prependTo(Vector<UChar, inlineCapacity>& result, unsigned pos, unsigned len) const
 613 {
 614     unsigned numberOfCharactersToCopy = std::min(len, length() - pos);
 615     if (!numberOfCharactersToCopy)
 616         return;
 617     if (is8Bit()) {
 618         size_t oldSize = result.size();
 619         result.resize(oldSize + numberOfCharactersToCopy);
 620         memmove(result.data() + numberOfCharactersToCopy, result.data(), oldSize * sizeof(UChar));
 621         StringImpl::copyChars(result.data(), m_impl->characters8() + pos, numberOfCharactersToCopy);
 622     } else {
 623         result.prepend(m_impl->characters16() + pos, numberOfCharactersToCopy);
 624     }
 625 }
 626
 627 // StringHash is the default hash for String
 628 template<typename T> struct DefaultHash;
 629 template<> struct DefaultHash<String> {
 630     typedef StringHash Hash;
 631 };
 632
 633 // Shared global empty string.
 634 WTF_EXPORT const String& emptyString();
 635 WTF_EXPORT const String& emptyString16Bit();
 636 WTF_EXPORT extern const String& xmlnsWithColon;
 637
 638 // Pretty printer for gtest. Declared here to avoid ODR violations.
 639 WTF_UNITTEST_HELPERS_EXPORT std::ostream& operator<<(std::ostream&, const String&);
 640
 641 } // namespace WTF
 642
 643 WTF_ALLOW_MOVE_AND_INIT_WITH_MEM_FUNCTIONS(String);
 644
 645 using WTF::CString;
 646 using WTF::KeepTrailingZeros;
 647 using WTF::StrictUTF8Conversion;
 648 using WTF::StrictUTF8ConversionReplacingUnpairedSurrogatesWithFFFD;
 649 using WTF::String;
 650 using WTF::emptyString;
 651 using WTF::emptyString16Bit;
 652 using WTF::append;
 653 using WTF::charactersAreAllASCII;
 654 using WTF::charactersToIntStrict;
 655 using WTF::charactersToUIntStrict;
 656 using WTF::charactersToInt64Strict;
 657 using WTF::charactersToUInt64Strict;
 658 using WTF::charactersToInt;
 659 using WTF::charactersToUInt;
 660 using WTF::charactersToInt64;
 661 using WTF::charactersToUInt64;
 662 using WTF::charactersToDouble;
 663 using WTF::charactersToFloat;
 664 using WTF::equal;
 665 using WTF::equalIgnoringCase;
 666 using WTF::find;
 667 using WTF::isAllSpecialCharacters;
 668 using WTF::isSpaceOrNewline;
 669 using WTF::reverseFind;
 670
 671 #include "wtf/text/AtomicString.h"
 672 #endif // WTFString_h