| OLD | NEW |
|---|
| (Empty) |
| 1 // Copyright 2013 The Chromium Authors. All rights reserved. | |
| 2 // Use of this source code is governed by a BSD-style license that can be | |
| 3 // found in the LICENSE file. | |
| 4 | |
| 5 #include "base/strings/string_util.h" | |
| 6 | |
| 7 #include <ctype.h> | |
| 8 #include <errno.h> | |
| 9 #include <math.h> | |
| 10 #include <stdarg.h> | |
| 11 #include <stdio.h> | |
| 12 #include <stdlib.h> | |
| 13 #include <string.h> | |
| 14 #include <time.h> | |
| 15 #include <wchar.h> | |
| 16 #include <wctype.h> | |
| 17 | |
| 18 #include <algorithm> | |
| 19 #include <vector> | |
| 20 | |
| 21 #include "base/basictypes.h" | |
| 22 #include "base/logging.h" | |
| 23 #include "base/memory/singleton.h" | |
| 24 #include "base/strings/string_split.h" | |
| 25 #include "base/strings/utf_string_conversion_utils.h" | |
| 26 #include "base/strings/utf_string_conversions.h" | |
| 27 #include "base/third_party/icu/icu_utf.h" | |
| 28 #include "build/build_config.h" | |
| 29 | |
| 30 namespace base { | |
| 31 | |
| 32 namespace { | |
| 33 | |
| 34 // Force the singleton used by EmptyString[16] to be a unique type. This | |
| 35 // prevents other code that might accidentally use Singleton<string> from | |
| 36 // getting our internal one. | |
| 37 struct EmptyStrings { | |
| 38 EmptyStrings() {} | |
| 39 const std::string s; | |
| 40 const string16 s16; | |
| 41 | |
| 42 static EmptyStrings* GetInstance() { | |
| 43 return Singleton<EmptyStrings>::get(); | |
| 44 } | |
| 45 }; | |
| 46 | |
| 47 // Used by ReplaceStringPlaceholders to track the position in the string of | |
| 48 // replaced parameters. | |
| 49 struct ReplacementOffset { | |
| 50 ReplacementOffset(uintptr_t parameter, size_t offset) | |
| 51 : parameter(parameter), | |
| 52 offset(offset) {} | |
| 53 | |
| 54 // Index of the parameter. | |
| 55 uintptr_t parameter; | |
| 56 | |
| 57 // Starting position in the string. | |
| 58 size_t offset; | |
| 59 }; | |
| 60 | |
| 61 static bool CompareParameter(const ReplacementOffset& elem1, | |
| 62 const ReplacementOffset& elem2) { | |
| 63 return elem1.parameter < elem2.parameter; | |
| 64 } | |
| 65 | |
| 66 // Assuming that a pointer is the size of a "machine word", then | |
| 67 // uintptr_t is an integer type that is also a machine word. | |
| 68 typedef uintptr_t MachineWord; | |
| 69 const uintptr_t kMachineWordAlignmentMask = sizeof(MachineWord) - 1; | |
| 70 | |
| 71 inline bool IsAlignedToMachineWord(const void* pointer) { | |
| 72 return !(reinterpret_cast<MachineWord>(pointer) & kMachineWordAlignmentMask); | |
| 73 } | |
| 74 | |
| 75 template<typename T> inline T* AlignToMachineWord(T* pointer) { | |
| 76 return reinterpret_cast<T*>(reinterpret_cast<MachineWord>(pointer) & | |
| 77 ~kMachineWordAlignmentMask); | |
| 78 } | |
| 79 | |
| 80 template<size_t size, typename CharacterType> struct NonASCIIMask; | |
| 81 template <> | |
| 82 struct NonASCIIMask<4, char16> { | |
| 83 static inline uint32_t value() { return 0xFF80FF80U; } | |
| 84 }; | |
| 85 template<> struct NonASCIIMask<4, char> { | |
| 86 static inline uint32_t value() { return 0x80808080U; } | |
| 87 }; | |
| 88 template <> | |
| 89 struct NonASCIIMask<8, char16> { | |
| 90 static inline uint64_t value() { return 0xFF80FF80FF80FF80ULL; } | |
| 91 }; | |
| 92 template<> struct NonASCIIMask<8, char> { | |
| 93 static inline uint64_t value() { return 0x8080808080808080ULL; } | |
| 94 }; | |
| 95 #if defined(WCHAR_T_IS_UTF32) | |
| 96 template<> struct NonASCIIMask<4, wchar_t> { | |
| 97 static inline uint32_t value() { return 0xFFFFFF80U; } | |
| 98 }; | |
| 99 template<> struct NonASCIIMask<8, wchar_t> { | |
| 100 static inline uint64_t value() { return 0xFFFFFF80FFFFFF80ULL; } | |
| 101 }; | |
| 102 #endif // WCHAR_T_IS_UTF32 | |
| 103 | |
| 104 // DO NOT USE. http://crbug.com/24917 | |
| 105 // | |
| 106 // tolower() will given incorrect results for non-ASCII characters. Use the | |
| 107 // ASCII version, base::i18n::ToLower, or base::i18n::FoldCase. This is here | |
| 108 // for backwards-compat for StartsWith until such calls can be updated. | |
| 109 struct CaseInsensitiveCompareDeprecated { | |
| 110 public: | |
| 111 bool operator()(char16 x, char16 y) const { return tolower(x) == tolower(y); } | |
| 112 }; | |
| 113 | |
| 114 } // namespace | |
| 115 | |
| 116 bool IsWprintfFormatPortable(const wchar_t* format) { | |
| 117 for (const wchar_t* position = format; *position != '\0'; ++position) { | |
| 118 if (*position == '%') { | |
| 119 bool in_specification = true; | |
| 120 bool modifier_l = false; | |
| 121 while (in_specification) { | |
| 122 // Eat up characters until reaching a known specifier. | |
| 123 if (*++position == '\0') { | |
| 124 // The format string ended in the middle of a specification. Call | |
| 125 // it portable because no unportable specifications were found. The | |
| 126 // string is equally broken on all platforms. | |
| 127 return true; | |
| 128 } | |
| 129 | |
| 130 if (*position == 'l') { | |
| 131 // 'l' is the only thing that can save the 's' and 'c' specifiers. | |
| 132 modifier_l = true; | |
| 133 } else if (((*position == 's' || *position == 'c') && !modifier_l) || | |
| 134 *position == 'S' || *position == 'C' || *position == 'F' || | |
| 135 *position == 'D' || *position == 'O' || *position == 'U') { | |
| 136 // Not portable. | |
| 137 return false; | |
| 138 } | |
| 139 | |
| 140 if (wcschr(L"diouxXeEfgGaAcspn%", *position)) { | |
| 141 // Portable, keep scanning the rest of the format string. | |
| 142 in_specification = false; | |
| 143 } | |
| 144 } | |
| 145 } | |
| 146 } | |
| 147 | |
| 148 return true; | |
| 149 } | |
| 150 | |
| 151 template <class StringType> | |
| 152 int CompareCaseInsensitiveASCIIT(BasicStringPiece<StringType> a, | |
| 153 BasicStringPiece<StringType> b) { | |
| 154 // Find the first characters that aren't equal and compare them. If the end | |
| 155 // of one of the strings is found before a nonequal character, the lengths | |
| 156 // of the strings are compared. | |
| 157 size_t i = 0; | |
| 158 while (i < a.length() && i < b.length()) { | |
| 159 typename StringType::value_type lower_a = ToLowerASCII(a[i]); | |
| 160 typename StringType::value_type lower_b = ToLowerASCII(b[i]); | |
| 161 if (lower_a < lower_b) | |
| 162 return -1; | |
| 163 if (lower_a > lower_b) | |
| 164 return 1; | |
| 165 i++; | |
| 166 } | |
| 167 | |
| 168 // End of one string hit before finding a different character. Expect the | |
| 169 // common case to be "strings equal" at this point so check that first. | |
| 170 if (a.length() == b.length()) | |
| 171 return 0; | |
| 172 | |
| 173 if (a.length() < b.length()) | |
| 174 return -1; | |
| 175 return 1; | |
| 176 } | |
| 177 | |
| 178 int CompareCaseInsensitiveASCII(StringPiece a, StringPiece b) { | |
| 179 return CompareCaseInsensitiveASCIIT<std::string>(a, b); | |
| 180 } | |
| 181 | |
| 182 int CompareCaseInsensitiveASCII(StringPiece16 a, StringPiece16 b) { | |
| 183 return CompareCaseInsensitiveASCIIT<string16>(a, b); | |
| 184 } | |
| 185 | |
| 186 bool EqualsCaseInsensitiveASCII(StringPiece a, StringPiece b) { | |
| 187 if (a.length() != b.length()) | |
| 188 return false; | |
| 189 return CompareCaseInsensitiveASCIIT<std::string>(a, b) == 0; | |
| 190 } | |
| 191 | |
| 192 bool EqualsCaseInsensitiveASCII(StringPiece16 a, StringPiece16 b) { | |
| 193 if (a.length() != b.length()) | |
| 194 return false; | |
| 195 return CompareCaseInsensitiveASCIIT<string16>(a, b) == 0; | |
| 196 } | |
| 197 | |
| 198 const std::string& EmptyString() { | |
| 199 return EmptyStrings::GetInstance()->s; | |
| 200 } | |
| 201 | |
| 202 const string16& EmptyString16() { | |
| 203 return EmptyStrings::GetInstance()->s16; | |
| 204 } | |
| 205 | |
| 206 template<typename STR> | |
| 207 bool ReplaceCharsT(const STR& input, | |
| 208 const STR& replace_chars, | |
| 209 const STR& replace_with, | |
| 210 STR* output) { | |
| 211 bool removed = false; | |
| 212 size_t replace_length = replace_with.length(); | |
| 213 | |
| 214 *output = input; | |
| 215 | |
| 216 size_t found = output->find_first_of(replace_chars); | |
| 217 while (found != STR::npos) { | |
| 218 removed = true; | |
| 219 output->replace(found, 1, replace_with); | |
| 220 found = output->find_first_of(replace_chars, found + replace_length); | |
| 221 } | |
| 222 | |
| 223 return removed; | |
| 224 } | |
| 225 | |
| 226 bool ReplaceChars(const string16& input, | |
| 227 const StringPiece16& replace_chars, | |
| 228 const string16& replace_with, | |
| 229 string16* output) { | |
| 230 return ReplaceCharsT(input, replace_chars.as_string(), replace_with, output); | |
| 231 } | |
| 232 | |
| 233 bool ReplaceChars(const std::string& input, | |
| 234 const StringPiece& replace_chars, | |
| 235 const std::string& replace_with, | |
| 236 std::string* output) { | |
| 237 return ReplaceCharsT(input, replace_chars.as_string(), replace_with, output); | |
| 238 } | |
| 239 | |
| 240 bool RemoveChars(const string16& input, | |
| 241 const StringPiece16& remove_chars, | |
| 242 string16* output) { | |
| 243 return ReplaceChars(input, remove_chars.as_string(), string16(), output); | |
| 244 } | |
| 245 | |
| 246 bool RemoveChars(const std::string& input, | |
| 247 const StringPiece& remove_chars, | |
| 248 std::string* output) { | |
| 249 return ReplaceChars(input, remove_chars.as_string(), std::string(), output); | |
| 250 } | |
| 251 | |
| 252 template <typename Str> | |
| 253 TrimPositions TrimStringT(const Str& input, | |
| 254 BasicStringPiece<Str> trim_chars, | |
| 255 TrimPositions positions, | |
| 256 Str* output) { | |
| 257 // Find the edges of leading/trailing whitespace as desired. Need to use | |
| 258 // a StringPiece version of input to be able to call find* on it with the | |
| 259 // StringPiece version of trim_chars (normally the trim_chars will be a | |
| 260 // constant so avoid making a copy). | |
| 261 BasicStringPiece<Str> input_piece(input); | |
| 262 const size_t last_char = input.length() - 1; | |
| 263 const size_t first_good_char = (positions & TRIM_LEADING) | |
| 264 ? input_piece.find_first_not_of(trim_chars) | |
| 265 : 0; | |
| 266 const size_t last_good_char = (positions & TRIM_TRAILING) | |
| 267 ? input_piece.find_last_not_of(trim_chars) | |
| 268 : last_char; | |
| 269 | |
| 270 // When the string was all trimmed, report that we stripped off characters | |
| 271 // from whichever position the caller was interested in. For empty input, we | |
| 272 // stripped no characters, but we still need to clear |output|. | |
| 273 if (input.empty() || (first_good_char == Str::npos) || | |
| 274 (last_good_char == Str::npos)) { | |
| 275 bool input_was_empty = input.empty(); // in case output == &input | |
| 276 output->clear(); | |
| 277 return input_was_empty ? TRIM_NONE : positions; | |
| 278 } | |
| 279 | |
| 280 // Trim. | |
| 281 *output = | |
| 282 input.substr(first_good_char, last_good_char - first_good_char + 1); | |
| 283 | |
| 284 // Return where we trimmed from. | |
| 285 return static_cast<TrimPositions>( | |
| 286 ((first_good_char == 0) ? TRIM_NONE : TRIM_LEADING) | | |
| 287 ((last_good_char == last_char) ? TRIM_NONE : TRIM_TRAILING)); | |
| 288 } | |
| 289 | |
| 290 bool TrimString(const string16& input, | |
| 291 StringPiece16 trim_chars, | |
| 292 string16* output) { | |
| 293 return TrimStringT(input, trim_chars, TRIM_ALL, output) != TRIM_NONE; | |
| 294 } | |
| 295 | |
| 296 bool TrimString(const std::string& input, | |
| 297 StringPiece trim_chars, | |
| 298 std::string* output) { | |
| 299 return TrimStringT(input, trim_chars, TRIM_ALL, output) != TRIM_NONE; | |
| 300 } | |
| 301 | |
| 302 template <typename Str> | |
| 303 BasicStringPiece<Str> TrimStringPieceT(BasicStringPiece<Str> input, | |
| 304 BasicStringPiece<Str> trim_chars, | |
| 305 TrimPositions positions) { | |
| 306 size_t begin = | |
| 307 (positions & TRIM_LEADING) ? input.find_first_not_of(trim_chars) : 0; | |
| 308 size_t end = (positions & TRIM_TRAILING) | |
| 309 ? input.find_last_not_of(trim_chars) + 1 | |
| 310 : input.size(); | |
| 311 return input.substr(begin, end - begin); | |
| 312 } | |
| 313 | |
| 314 StringPiece16 TrimString(StringPiece16 input, | |
| 315 const StringPiece16& trim_chars, | |
| 316 TrimPositions positions) { | |
| 317 return TrimStringPieceT(input, trim_chars, positions); | |
| 318 } | |
| 319 | |
| 320 StringPiece TrimString(StringPiece input, | |
| 321 const StringPiece& trim_chars, | |
| 322 TrimPositions positions) { | |
| 323 return TrimStringPieceT(input, trim_chars, positions); | |
| 324 } | |
| 325 | |
| 326 void TruncateUTF8ToByteSize(const std::string& input, | |
| 327 const size_t byte_size, | |
| 328 std::string* output) { | |
| 329 DCHECK(output); | |
| 330 if (byte_size > input.length()) { | |
| 331 *output = input; | |
| 332 return; | |
| 333 } | |
| 334 DCHECK_LE(byte_size, static_cast<uint32>(kint32max)); | |
| 335 // Note: This cast is necessary because CBU8_NEXT uses int32s. | |
| 336 int32 truncation_length = static_cast<int32>(byte_size); | |
| 337 int32 char_index = truncation_length - 1; | |
| 338 const char* data = input.data(); | |
| 339 | |
| 340 // Using CBU8, we will move backwards from the truncation point | |
| 341 // to the beginning of the string looking for a valid UTF8 | |
| 342 // character. Once a full UTF8 character is found, we will | |
| 343 // truncate the string to the end of that character. | |
| 344 while (char_index >= 0) { | |
| 345 int32 prev = char_index; | |
| 346 base_icu::UChar32 code_point = 0; | |
| 347 CBU8_NEXT(data, char_index, truncation_length, code_point); | |
| 348 if (!IsValidCharacter(code_point) || | |
| 349 !IsValidCodepoint(code_point)) { | |
| 350 char_index = prev - 1; | |
| 351 } else { | |
| 352 break; | |
| 353 } | |
| 354 } | |
| 355 | |
| 356 if (char_index >= 0 ) | |
| 357 *output = input.substr(0, char_index); | |
| 358 else | |
| 359 output->clear(); | |
| 360 } | |
| 361 | |
| 362 TrimPositions TrimWhitespace(const string16& input, | |
| 363 TrimPositions positions, | |
| 364 string16* output) { | |
| 365 return TrimStringT(input, StringPiece16(kWhitespaceUTF16), positions, output); | |
| 366 } | |
| 367 | |
| 368 StringPiece16 TrimWhitespaceASCII(StringPiece16 input, | |
| 369 TrimPositions positions) { | |
| 370 return TrimStringPieceT(input, StringPiece16(kWhitespaceUTF16), positions); | |
| 371 } | |
| 372 | |
| 373 TrimPositions TrimWhitespaceASCII(const std::string& input, | |
| 374 TrimPositions positions, | |
| 375 std::string* output) { | |
| 376 return TrimStringT(input, StringPiece(kWhitespaceASCII), positions, output); | |
| 377 } | |
| 378 | |
| 379 StringPiece TrimWhitespaceASCII(StringPiece input, TrimPositions positions) { | |
| 380 return TrimStringPieceT(input, StringPiece(kWhitespaceASCII), positions); | |
| 381 } | |
| 382 | |
| 383 // This function is only for backward-compatibility. | |
| 384 // To be removed when all callers are updated. | |
| 385 TrimPositions TrimWhitespace(const std::string& input, | |
| 386 TrimPositions positions, | |
| 387 std::string* output) { | |
| 388 return TrimWhitespaceASCII(input, positions, output); | |
| 389 } | |
| 390 | |
| 391 template<typename STR> | |
| 392 STR CollapseWhitespaceT(const STR& text, | |
| 393 bool trim_sequences_with_line_breaks) { | |
| 394 STR result; | |
| 395 result.resize(text.size()); | |
| 396 | |
| 397 // Set flags to pretend we're already in a trimmed whitespace sequence, so we | |
| 398 // will trim any leading whitespace. | |
| 399 bool in_whitespace = true; | |
| 400 bool already_trimmed = true; | |
| 401 | |
| 402 int chars_written = 0; | |
| 403 for (typename STR::const_iterator i(text.begin()); i != text.end(); ++i) { | |
| 404 if (IsUnicodeWhitespace(*i)) { | |
| 405 if (!in_whitespace) { | |
| 406 // Reduce all whitespace sequences to a single space. | |
| 407 in_whitespace = true; | |
| 408 result[chars_written++] = L' '; | |
| 409 } | |
| 410 if (trim_sequences_with_line_breaks && !already_trimmed && | |
| 411 ((*i == '\n') || (*i == '\r'))) { | |
| 412 // Whitespace sequences containing CR or LF are eliminated entirely. | |
| 413 already_trimmed = true; | |
| 414 --chars_written; | |
| 415 } | |
| 416 } else { | |
| 417 // Non-whitespace chracters are copied straight across. | |
| 418 in_whitespace = false; | |
| 419 already_trimmed = false; | |
| 420 result[chars_written++] = *i; | |
| 421 } | |
| 422 } | |
| 423 | |
| 424 if (in_whitespace && !already_trimmed) { | |
| 425 // Any trailing whitespace is eliminated. | |
| 426 --chars_written; | |
| 427 } | |
| 428 | |
| 429 result.resize(chars_written); | |
| 430 return result; | |
| 431 } | |
| 432 | |
| 433 string16 CollapseWhitespace(const string16& text, | |
| 434 bool trim_sequences_with_line_breaks) { | |
| 435 return CollapseWhitespaceT(text, trim_sequences_with_line_breaks); | |
| 436 } | |
| 437 | |
| 438 std::string CollapseWhitespaceASCII(const std::string& text, | |
| 439 bool trim_sequences_with_line_breaks) { | |
| 440 return CollapseWhitespaceT(text, trim_sequences_with_line_breaks); | |
| 441 } | |
| 442 | |
| 443 bool ContainsOnlyChars(const StringPiece& input, | |
| 444 const StringPiece& characters) { | |
| 445 return input.find_first_not_of(characters) == StringPiece::npos; | |
| 446 } | |
| 447 | |
| 448 bool ContainsOnlyChars(const StringPiece16& input, | |
| 449 const StringPiece16& characters) { | |
| 450 return input.find_first_not_of(characters) == StringPiece16::npos; | |
| 451 } | |
| 452 | |
| 453 template <class Char> | |
| 454 inline bool DoIsStringASCII(const Char* characters, size_t length) { | |
| 455 MachineWord all_char_bits = 0; | |
| 456 const Char* end = characters + length; | |
| 457 | |
| 458 // Prologue: align the input. | |
| 459 while (!IsAlignedToMachineWord(characters) && characters != end) { | |
| 460 all_char_bits |= *characters; | |
| 461 ++characters; | |
| 462 } | |
| 463 | |
| 464 // Compare the values of CPU word size. | |
| 465 const Char* word_end = AlignToMachineWord(end); | |
| 466 const size_t loop_increment = sizeof(MachineWord) / sizeof(Char); | |
| 467 while (characters < word_end) { | |
| 468 all_char_bits |= *(reinterpret_cast<const MachineWord*>(characters)); | |
| 469 characters += loop_increment; | |
| 470 } | |
| 471 | |
| 472 // Process the remaining bytes. | |
| 473 while (characters != end) { | |
| 474 all_char_bits |= *characters; | |
| 475 ++characters; | |
| 476 } | |
| 477 | |
| 478 MachineWord non_ascii_bit_mask = | |
| 479 NonASCIIMask<sizeof(MachineWord), Char>::value(); | |
| 480 return !(all_char_bits & non_ascii_bit_mask); | |
| 481 } | |
| 482 | |
| 483 bool IsStringASCII(const StringPiece& str) { | |
| 484 return DoIsStringASCII(str.data(), str.length()); | |
| 485 } | |
| 486 | |
| 487 bool IsStringASCII(const StringPiece16& str) { | |
| 488 return DoIsStringASCII(str.data(), str.length()); | |
| 489 } | |
| 490 | |
| 491 bool IsStringASCII(const string16& str) { | |
| 492 return DoIsStringASCII(str.data(), str.length()); | |
| 493 } | |
| 494 | |
| 495 #if defined(WCHAR_T_IS_UTF32) | |
| 496 bool IsStringASCII(const std::wstring& str) { | |
| 497 return DoIsStringASCII(str.data(), str.length()); | |
| 498 } | |
| 499 #endif | |
| 500 | |
| 501 bool IsStringUTF8(const StringPiece& str) { | |
| 502 const char *src = str.data(); | |
| 503 int32 src_len = static_cast<int32>(str.length()); | |
| 504 int32 char_index = 0; | |
| 505 | |
| 506 while (char_index < src_len) { | |
| 507 int32 code_point; | |
| 508 CBU8_NEXT(src, char_index, src_len, code_point); | |
| 509 if (!IsValidCharacter(code_point)) | |
| 510 return false; | |
| 511 } | |
| 512 return true; | |
| 513 } | |
| 514 | |
| 515 template<typename Iter> | |
| 516 static inline bool DoLowerCaseEqualsASCII(Iter a_begin, | |
| 517 Iter a_end, | |
| 518 const char* b) { | |
| 519 for (Iter it = a_begin; it != a_end; ++it, ++b) { | |
| 520 if (!*b || ToLowerASCII(*it) != *b) | |
| 521 return false; | |
| 522 } | |
| 523 return *b == 0; | |
| 524 } | |
| 525 | |
| 526 // Front-ends for LowerCaseEqualsASCII. | |
| 527 bool LowerCaseEqualsASCII(const std::string& a, const char* b) { | |
| 528 return DoLowerCaseEqualsASCII(a.begin(), a.end(), b); | |
| 529 } | |
| 530 | |
| 531 bool LowerCaseEqualsASCII(const string16& a, const char* b) { | |
| 532 return DoLowerCaseEqualsASCII(a.begin(), a.end(), b); | |
| 533 } | |
| 534 | |
| 535 bool LowerCaseEqualsASCII(std::string::const_iterator a_begin, | |
| 536 std::string::const_iterator a_end, | |
| 537 const char* b) { | |
| 538 return DoLowerCaseEqualsASCII(a_begin, a_end, b); | |
| 539 } | |
| 540 | |
| 541 bool LowerCaseEqualsASCII(string16::const_iterator a_begin, | |
| 542 string16::const_iterator a_end, | |
| 543 const char* b) { | |
| 544 return DoLowerCaseEqualsASCII(a_begin, a_end, b); | |
| 545 } | |
| 546 | |
| 547 bool LowerCaseEqualsASCII(const char* a_begin, | |
| 548 const char* a_end, | |
| 549 const char* b) { | |
| 550 return DoLowerCaseEqualsASCII(a_begin, a_end, b); | |
| 551 } | |
| 552 | |
| 553 bool LowerCaseEqualsASCII(const char* a_begin, | |
| 554 const char* a_end, | |
| 555 const char* b_begin, | |
| 556 const char* b_end) { | |
| 557 while (a_begin != a_end && b_begin != b_end && | |
| 558 ToLowerASCII(*a_begin) == *b_begin) { | |
| 559 a_begin++; | |
| 560 b_begin++; | |
| 561 } | |
| 562 return a_begin == a_end && b_begin == b_end; | |
| 563 } | |
| 564 | |
| 565 bool LowerCaseEqualsASCII(const char16* a_begin, | |
| 566 const char16* a_end, | |
| 567 const char* b) { | |
| 568 return DoLowerCaseEqualsASCII(a_begin, a_end, b); | |
| 569 } | |
| 570 | |
| 571 bool EqualsASCII(const string16& a, const StringPiece& b) { | |
| 572 if (a.length() != b.length()) | |
| 573 return false; | |
| 574 return std::equal(b.begin(), b.end(), a.begin()); | |
| 575 } | |
| 576 | |
| 577 template <typename Str> | |
| 578 bool StartsWithT(BasicStringPiece<Str> str, | |
| 579 BasicStringPiece<Str> search_for, | |
| 580 CompareCase case_sensitivity) { | |
| 581 if (search_for.size() > str.size()) | |
| 582 return false; | |
| 583 | |
| 584 BasicStringPiece<Str> source = str.substr(0, search_for.size()); | |
| 585 | |
| 586 switch (case_sensitivity) { | |
| 587 case CompareCase::SENSITIVE: | |
| 588 return source == search_for; | |
| 589 | |
| 590 case CompareCase::INSENSITIVE_ASCII: | |
| 591 return std::equal( | |
| 592 search_for.begin(), search_for.end(), source.begin(), | |
| 593 CaseInsensitiveCompareASCII<typename Str::value_type>()); | |
| 594 | |
| 595 default: | |
| 596 NOTREACHED(); | |
| 597 return false; | |
| 598 } | |
| 599 } | |
| 600 | |
| 601 bool StartsWith(StringPiece str, | |
| 602 StringPiece search_for, | |
| 603 CompareCase case_sensitivity) { | |
| 604 return StartsWithT<std::string>(str, search_for, case_sensitivity); | |
| 605 } | |
| 606 | |
| 607 bool StartsWith(StringPiece16 str, | |
| 608 StringPiece16 search_for, | |
| 609 CompareCase case_sensitivity) { | |
| 610 return StartsWithT<string16>(str, search_for, case_sensitivity); | |
| 611 } | |
| 612 | |
| 613 bool StartsWith(const string16& str, | |
| 614 const string16& search, | |
| 615 bool case_sensitive) { | |
| 616 if (!case_sensitive) { | |
| 617 // This function was originally written using the current locale functions | |
| 618 // for case-insensitive comparisons. Emulate this behavior until callers | |
| 619 // can be converted either to use the case-insensitive ASCII one (most | |
| 620 // callers) or ICU functions in base_i18n. | |
| 621 if (search.size() > str.size()) | |
| 622 return false; | |
| 623 return std::equal(search.begin(), search.end(), str.begin(), | |
| 624 CaseInsensitiveCompareDeprecated()); | |
| 625 } | |
| 626 return StartsWith(StringPiece16(str), StringPiece16(search), | |
| 627 CompareCase::SENSITIVE); | |
| 628 } | |
| 629 | |
| 630 template <typename Str> | |
| 631 bool EndsWithT(BasicStringPiece<Str> str, | |
| 632 BasicStringPiece<Str> search_for, | |
| 633 CompareCase case_sensitivity) { | |
| 634 if (search_for.size() > str.size()) | |
| 635 return false; | |
| 636 | |
| 637 BasicStringPiece<Str> source = | |
| 638 str.substr(str.size() - search_for.size(), search_for.size()); | |
| 639 | |
| 640 switch (case_sensitivity) { | |
| 641 case CompareCase::SENSITIVE: | |
| 642 return source == search_for; | |
| 643 | |
| 644 case CompareCase::INSENSITIVE_ASCII: | |
| 645 return std::equal( | |
| 646 source.begin(), source.end(), search_for.begin(), | |
| 647 CaseInsensitiveCompareASCII<typename Str::value_type>()); | |
| 648 | |
| 649 default: | |
| 650 NOTREACHED(); | |
| 651 return false; | |
| 652 } | |
| 653 } | |
| 654 | |
| 655 bool EndsWith(StringPiece str, | |
| 656 StringPiece search_for, | |
| 657 CompareCase case_sensitivity) { | |
| 658 return EndsWithT<std::string>(str, search_for, case_sensitivity); | |
| 659 } | |
| 660 | |
| 661 bool EndsWith(StringPiece16 str, | |
| 662 StringPiece16 search_for, | |
| 663 CompareCase case_sensitivity) { | |
| 664 return EndsWithT<string16>(str, search_for, case_sensitivity); | |
| 665 } | |
| 666 | |
| 667 bool EndsWith(const string16& str, | |
| 668 const string16& search, | |
| 669 bool case_sensitive) { | |
| 670 if (!case_sensitive) { | |
| 671 // This function was originally written using the current locale functions | |
| 672 // for case-insensitive comparisons. Emulate this behavior until callers | |
| 673 // can be converted either to use the case-insensitive ASCII one (most | |
| 674 // callers) or ICU functions in base_i18n. | |
| 675 if (search.size() > str.size()) | |
| 676 return false; | |
| 677 return std::equal(search.begin(), search.end(), | |
| 678 str.begin() + (str.size() - search.size()), | |
| 679 CaseInsensitiveCompareDeprecated()); | |
| 680 } | |
| 681 return EndsWith(StringPiece16(str), StringPiece16(search), | |
| 682 CompareCase::SENSITIVE); | |
| 683 } | |
| 684 | |
| 685 char HexDigitToInt(wchar_t c) { | |
| 686 DCHECK(IsHexDigit(c)); | |
| 687 if (c >= '0' && c <= '9') | |
| 688 return static_cast<char>(c - '0'); | |
| 689 if (c >= 'A' && c <= 'F') | |
| 690 return static_cast<char>(c - 'A' + 10); | |
| 691 if (c >= 'a' && c <= 'f') | |
| 692 return static_cast<char>(c - 'a' + 10); | |
| 693 return 0; | |
| 694 } | |
| 695 | |
| 696 static const char* const kByteStringsUnlocalized[] = { | |
| 697 " B", | |
| 698 " kB", | |
| 699 " MB", | |
| 700 " GB", | |
| 701 " TB", | |
| 702 " PB" | |
| 703 }; | |
| 704 | |
| 705 string16 FormatBytesUnlocalized(int64 bytes) { | |
| 706 double unit_amount = static_cast<double>(bytes); | |
| 707 size_t dimension = 0; | |
| 708 const int kKilo = 1024; | |
| 709 while (unit_amount >= kKilo && | |
| 710 dimension < arraysize(kByteStringsUnlocalized) - 1) { | |
| 711 unit_amount /= kKilo; | |
| 712 dimension++; | |
| 713 } | |
| 714 | |
| 715 char buf[64]; | |
| 716 if (bytes != 0 && dimension > 0 && unit_amount < 100) { | |
| 717 base::snprintf(buf, arraysize(buf), "%.1lf%s", unit_amount, | |
| 718 kByteStringsUnlocalized[dimension]); | |
| 719 } else { | |
| 720 base::snprintf(buf, arraysize(buf), "%.0lf%s", unit_amount, | |
| 721 kByteStringsUnlocalized[dimension]); | |
| 722 } | |
| 723 | |
| 724 return ASCIIToUTF16(buf); | |
| 725 } | |
| 726 | |
| 727 // Runs in O(n) time in the length of |str|. | |
| 728 template <class StringType> | |
| 729 void DoReplaceSubstringsAfterOffset(StringType* str, | |
| 730 size_t offset, | |
| 731 BasicStringPiece<StringType> find_this, | |
| 732 BasicStringPiece<StringType> replace_with, | |
| 733 bool replace_all) { | |
| 734 DCHECK(!find_this.empty()); | |
| 735 | |
| 736 // If the find string doesn't appear, there's nothing to do. | |
| 737 offset = str->find(find_this.data(), offset, find_this.size()); | |
| 738 if (offset == StringType::npos) | |
| 739 return; | |
| 740 | |
| 741 // If we're only replacing one instance, there's no need to do anything | |
| 742 // complicated. | |
| 743 size_t find_length = find_this.length(); | |
| 744 if (!replace_all) { | |
| 745 str->replace(offset, find_length, replace_with.data(), replace_with.size()); | |
| 746 return; | |
| 747 } | |
| 748 | |
| 749 // If the find and replace strings are the same length, we can simply use | |
| 750 // replace() on each instance, and finish the entire operation in O(n) time. | |
| 751 size_t replace_length = replace_with.length(); | |
| 752 if (find_length == replace_length) { | |
| 753 do { | |
| 754 str->replace(offset, find_length, replace_with.data(), | |
| 755 replace_with.size()); | |
| 756 offset = str->find(find_this.data(), offset + replace_length, | |
| 757 find_this.size()); | |
| 758 } while (offset != StringType::npos); | |
| 759 return; | |
| 760 } | |
| 761 | |
| 762 // Since the find and replace strings aren't the same length, a loop like the | |
| 763 // one above would be O(n^2) in the worst case, as replace() will shift the | |
| 764 // entire remaining string each time. We need to be more clever to keep | |
| 765 // things O(n). | |
| 766 // | |
| 767 // If we're shortening the string, we can alternate replacements with shifting | |
| 768 // forward the intervening characters using memmove(). | |
| 769 size_t str_length = str->length(); | |
| 770 if (find_length > replace_length) { | |
| 771 size_t write_offset = offset; | |
| 772 do { | |
| 773 if (replace_length) { | |
| 774 str->replace(write_offset, replace_length, replace_with.data(), | |
| 775 replace_with.size()); | |
| 776 write_offset += replace_length; | |
| 777 } | |
| 778 size_t read_offset = offset + find_length; | |
| 779 offset = | |
| 780 std::min(str->find(find_this.data(), read_offset, find_this.size()), | |
| 781 str_length); | |
| 782 size_t length = offset - read_offset; | |
| 783 if (length) { | |
| 784 memmove(&(*str)[write_offset], &(*str)[read_offset], | |
| 785 length * sizeof(typename StringType::value_type)); | |
| 786 write_offset += length; | |
| 787 } | |
| 788 } while (offset < str_length); | |
| 789 str->resize(write_offset); | |
| 790 return; | |
| 791 } | |
| 792 | |
| 793 // We're lengthening the string. We can use alternating replacements and | |
| 794 // memmove() calls like above, but we need to precalculate the final string | |
| 795 // length and then expand from back-to-front to avoid overwriting the string | |
| 796 // as we're reading it, needing to shift, or having to copy to a second string | |
| 797 // temporarily. | |
| 798 size_t first_match = offset; | |
| 799 | |
| 800 // First, calculate the final length and resize the string. | |
| 801 size_t final_length = str_length; | |
| 802 size_t expansion = replace_length - find_length; | |
| 803 size_t current_match; | |
| 804 do { | |
| 805 final_length += expansion; | |
| 806 // Minor optimization: save this offset into |current_match|, so that on | |
| 807 // exit from the loop, |current_match| will point at the last instance of | |
| 808 // the find string, and we won't need to find() it again immediately. | |
| 809 current_match = offset; | |
| 810 offset = | |
| 811 str->find(find_this.data(), offset + find_length, find_this.size()); | |
| 812 } while (offset != StringType::npos); | |
| 813 str->resize(final_length); | |
| 814 | |
| 815 // Now do the replacement loop, working backwards through the string. | |
| 816 for (size_t prev_match = str_length, write_offset = final_length;; | |
| 817 current_match = | |
| 818 str->rfind(find_this.data(), current_match - 1, find_this.size())) { | |
| 819 size_t read_offset = current_match + find_length; | |
| 820 size_t length = prev_match - read_offset; | |
| 821 if (length) { | |
| 822 write_offset -= length; | |
| 823 memmove(&(*str)[write_offset], &(*str)[read_offset], | |
| 824 length * sizeof(typename StringType::value_type)); | |
| 825 } | |
| 826 write_offset -= replace_length; | |
| 827 str->replace(write_offset, replace_length, replace_with.data(), | |
| 828 replace_with.size()); | |
| 829 if (current_match == first_match) | |
| 830 return; | |
| 831 prev_match = current_match; | |
| 832 } | |
| 833 } | |
| 834 | |
| 835 void ReplaceFirstSubstringAfterOffset(string16* str, | |
| 836 size_t start_offset, | |
| 837 StringPiece16 find_this, | |
| 838 StringPiece16 replace_with) { | |
| 839 DoReplaceSubstringsAfterOffset<string16>( | |
| 840 str, start_offset, find_this, replace_with, false); // Replace first. | |
| 841 } | |
| 842 | |
| 843 void ReplaceFirstSubstringAfterOffset(std::string* str, | |
| 844 size_t start_offset, | |
| 845 StringPiece find_this, | |
| 846 StringPiece replace_with) { | |
| 847 DoReplaceSubstringsAfterOffset<std::string>( | |
| 848 str, start_offset, find_this, replace_with, false); // Replace first. | |
| 849 } | |
| 850 | |
| 851 void ReplaceSubstringsAfterOffset(string16* str, | |
| 852 size_t start_offset, | |
| 853 StringPiece16 find_this, | |
| 854 StringPiece16 replace_with) { | |
| 855 DoReplaceSubstringsAfterOffset<string16>(str, start_offset, find_this, | |
| 856 replace_with, true); // Replace all. | |
| 857 } | |
| 858 | |
| 859 void ReplaceSubstringsAfterOffset(std::string* str, | |
| 860 size_t start_offset, | |
| 861 StringPiece find_this, | |
| 862 StringPiece replace_with) { | |
| 863 DoReplaceSubstringsAfterOffset<std::string>( | |
| 864 str, start_offset, find_this, replace_with, true); // Replace all. | |
| 865 } | |
| 866 | |
| 867 template <class string_type> | |
| 868 inline typename string_type::value_type* WriteIntoT(string_type* str, | |
| 869 size_t length_with_null) { | |
| 870 DCHECK_GT(length_with_null, 1u); | |
| 871 str->reserve(length_with_null); | |
| 872 str->resize(length_with_null - 1); | |
| 873 return &((*str)[0]); | |
| 874 } | |
| 875 | |
| 876 char* WriteInto(std::string* str, size_t length_with_null) { | |
| 877 return WriteIntoT(str, length_with_null); | |
| 878 } | |
| 879 | |
| 880 char16* WriteInto(string16* str, size_t length_with_null) { | |
| 881 return WriteIntoT(str, length_with_null); | |
| 882 } | |
| 883 | |
| 884 template <typename STR> | |
| 885 static STR JoinStringT(const std::vector<STR>& parts, | |
| 886 BasicStringPiece<STR> sep) { | |
| 887 if (parts.empty()) | |
| 888 return STR(); | |
| 889 | |
| 890 STR result(parts[0]); | |
| 891 auto iter = parts.begin(); | |
| 892 ++iter; | |
| 893 | |
| 894 for (; iter != parts.end(); ++iter) { | |
| 895 sep.AppendToString(&result); | |
| 896 result += *iter; | |
| 897 } | |
| 898 | |
| 899 return result; | |
| 900 } | |
| 901 | |
| 902 std::string JoinString(const std::vector<std::string>& parts, | |
| 903 StringPiece separator) { | |
| 904 return JoinStringT(parts, separator); | |
| 905 } | |
| 906 | |
| 907 string16 JoinString(const std::vector<string16>& parts, | |
| 908 StringPiece16 separator) { | |
| 909 return JoinStringT(parts, separator); | |
| 910 } | |
| 911 | |
| 912 template <class FormatStringType, class OutStringType> | |
| 913 OutStringType DoReplaceStringPlaceholders( | |
| 914 const FormatStringType& format_string, | |
| 915 const std::vector<OutStringType>& subst, | |
| 916 std::vector<size_t>* offsets) { | |
| 917 size_t substitutions = subst.size(); | |
| 918 | |
| 919 size_t sub_length = 0; | |
| 920 for (const auto& cur : subst) | |
| 921 sub_length += cur.length(); | |
| 922 | |
| 923 OutStringType formatted; | |
| 924 formatted.reserve(format_string.length() + sub_length); | |
| 925 | |
| 926 std::vector<ReplacementOffset> r_offsets; | |
| 927 for (auto i = format_string.begin(); i != format_string.end(); ++i) { | |
| 928 if ('$' == *i) { | |
| 929 if (i + 1 != format_string.end()) { | |
| 930 ++i; | |
| 931 DCHECK('$' == *i || '1' <= *i) << "Invalid placeholder: " << *i; | |
| 932 if ('$' == *i) { | |
| 933 while (i != format_string.end() && '$' == *i) { | |
| 934 formatted.push_back('$'); | |
| 935 ++i; | |
| 936 } | |
| 937 --i; | |
| 938 } else { | |
| 939 uintptr_t index = 0; | |
| 940 while (i != format_string.end() && '0' <= *i && *i <= '9') { | |
| 941 index *= 10; | |
| 942 index += *i - '0'; | |
| 943 ++i; | |
| 944 } | |
| 945 --i; | |
| 946 index -= 1; | |
| 947 if (offsets) { | |
| 948 ReplacementOffset r_offset(index, | |
| 949 static_cast<int>(formatted.size())); | |
| 950 r_offsets.insert(std::lower_bound(r_offsets.begin(), | |
| 951 r_offsets.end(), | |
| 952 r_offset, | |
| 953 &CompareParameter), | |
| 954 r_offset); | |
| 955 } | |
| 956 if (index < substitutions) | |
| 957 formatted.append(subst.at(index)); | |
| 958 } | |
| 959 } | |
| 960 } else { | |
| 961 formatted.push_back(*i); | |
| 962 } | |
| 963 } | |
| 964 if (offsets) { | |
| 965 for (const auto& cur : r_offsets) | |
| 966 offsets->push_back(cur.offset); | |
| 967 } | |
| 968 return formatted; | |
| 969 } | |
| 970 | |
| 971 string16 ReplaceStringPlaceholders(const string16& format_string, | |
| 972 const std::vector<string16>& subst, | |
| 973 std::vector<size_t>* offsets) { | |
| 974 return DoReplaceStringPlaceholders(format_string, subst, offsets); | |
| 975 } | |
| 976 | |
| 977 std::string ReplaceStringPlaceholders(const StringPiece& format_string, | |
| 978 const std::vector<std::string>& subst, | |
| 979 std::vector<size_t>* offsets) { | |
| 980 return DoReplaceStringPlaceholders(format_string, subst, offsets); | |
| 981 } | |
| 982 | |
| 983 string16 ReplaceStringPlaceholders(const string16& format_string, | |
| 984 const string16& a, | |
| 985 size_t* offset) { | |
| 986 std::vector<size_t> offsets; | |
| 987 std::vector<string16> subst; | |
| 988 subst.push_back(a); | |
| 989 string16 result = ReplaceStringPlaceholders(format_string, subst, &offsets); | |
| 990 | |
| 991 DCHECK_EQ(1U, offsets.size()); | |
| 992 if (offset) | |
| 993 *offset = offsets[0]; | |
| 994 return result; | |
| 995 } | |
| 996 | |
| 997 // The following code is compatible with the OpenBSD lcpy interface. See: | |
| 998 // http://www.gratisoft.us/todd/papers/strlcpy.html | |
| 999 // ftp://ftp.openbsd.org/pub/OpenBSD/src/lib/libc/string/{wcs,str}lcpy.c | |
| 1000 | |
| 1001 namespace { | |
| 1002 | |
| 1003 template <typename CHAR> | |
| 1004 size_t lcpyT(CHAR* dst, const CHAR* src, size_t dst_size) { | |
| 1005 for (size_t i = 0; i < dst_size; ++i) { | |
| 1006 if ((dst[i] = src[i]) == 0) // We hit and copied the terminating NULL. | |
| 1007 return i; | |
| 1008 } | |
| 1009 | |
| 1010 // We were left off at dst_size. We over copied 1 byte. Null terminate. | |
| 1011 if (dst_size != 0) | |
| 1012 dst[dst_size - 1] = 0; | |
| 1013 | |
| 1014 // Count the rest of the |src|, and return it's length in characters. | |
| 1015 while (src[dst_size]) ++dst_size; | |
| 1016 return dst_size; | |
| 1017 } | |
| 1018 | |
| 1019 } // namespace | |
| 1020 | |
| 1021 size_t strlcpy(char* dst, const char* src, size_t dst_size) { | |
| 1022 return lcpyT<char>(dst, src, dst_size); | |
| 1023 } | |
| 1024 size_t wcslcpy(wchar_t* dst, const wchar_t* src, size_t dst_size) { | |
| 1025 return lcpyT<wchar_t>(dst, src, dst_size); | |
| 1026 } | |
| 1027 | |
| 1028 } // namespace base | |
| OLD | NEW |
|---|
Chromium Code Reviews
