1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/security/sandbox/chromium/base/strings/utf_string_conversion_utils.cc Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,148 @@
1.4 +// Copyright (c) 2009 The Chromium Authors. All rights reserved.
1.5 +// Use of this source code is governed by a BSD-style license that can be
1.6 +// found in the LICENSE file.
1.7 +
1.8 +#include "base/strings/utf_string_conversion_utils.h"
1.9 +
1.10 +#include "base/third_party/icu/icu_utf.h"
1.11 +
1.12 +namespace base {
1.13 +
1.14 +// ReadUnicodeCharacter --------------------------------------------------------
1.15 +
1.16 +bool ReadUnicodeCharacter(const char* src,
1.17 + int32 src_len,
1.18 + int32* char_index,
1.19 + uint32* code_point_out) {
1.20 + // U8_NEXT expects to be able to use -1 to signal an error, so we must
1.21 + // use a signed type for code_point. But this function returns false
1.22 + // on error anyway, so code_point_out is unsigned.
1.23 + int32 code_point;
1.24 + CBU8_NEXT(src, *char_index, src_len, code_point);
1.25 + *code_point_out = static_cast<uint32>(code_point);
1.26 +
1.27 + // The ICU macro above moves to the next char, we want to point to the last
1.28 + // char consumed.
1.29 + (*char_index)--;
1.30 +
1.31 + // Validate the decoded value.
1.32 + return IsValidCodepoint(code_point);
1.33 +}
1.34 +
1.35 +bool ReadUnicodeCharacter(const char16* src,
1.36 + int32 src_len,
1.37 + int32* char_index,
1.38 + uint32* code_point) {
1.39 + if (CBU16_IS_SURROGATE(src[*char_index])) {
1.40 + if (!CBU16_IS_SURROGATE_LEAD(src[*char_index]) ||
1.41 + *char_index + 1 >= src_len ||
1.42 + !CBU16_IS_TRAIL(src[*char_index + 1])) {
1.43 + // Invalid surrogate pair.
1.44 + return false;
1.45 + }
1.46 +
1.47 + // Valid surrogate pair.
1.48 + *code_point = CBU16_GET_SUPPLEMENTARY(src[*char_index],
1.49 + src[*char_index + 1]);
1.50 + (*char_index)++;
1.51 + } else {
1.52 + // Not a surrogate, just one 16-bit word.
1.53 + *code_point = src[*char_index];
1.54 + }
1.55 +
1.56 + return IsValidCodepoint(*code_point);
1.57 +}
1.58 +
1.59 +#if defined(WCHAR_T_IS_UTF32)
1.60 +bool ReadUnicodeCharacter(const wchar_t* src,
1.61 + int32 src_len,
1.62 + int32* char_index,
1.63 + uint32* code_point) {
1.64 + // Conversion is easy since the source is 32-bit.
1.65 + *code_point = src[*char_index];
1.66 +
1.67 + // Validate the value.
1.68 + return IsValidCodepoint(*code_point);
1.69 +}
1.70 +#endif // defined(WCHAR_T_IS_UTF32)
1.71 +
1.72 +// WriteUnicodeCharacter -------------------------------------------------------
1.73 +
1.74 +size_t WriteUnicodeCharacter(uint32 code_point, std::string* output) {
1.75 + if (code_point <= 0x7f) {
1.76 + // Fast path the common case of one byte.
1.77 + output->push_back(code_point);
1.78 + return 1;
1.79 + }
1.80 +
1.81 +
1.82 + // CBU8_APPEND_UNSAFE can append up to 4 bytes.
1.83 + size_t char_offset = output->length();
1.84 + size_t original_char_offset = char_offset;
1.85 + output->resize(char_offset + CBU8_MAX_LENGTH);
1.86 +
1.87 + CBU8_APPEND_UNSAFE(&(*output)[0], char_offset, code_point);
1.88 +
1.89 + // CBU8_APPEND_UNSAFE will advance our pointer past the inserted character, so
1.90 + // it will represent the new length of the string.
1.91 + output->resize(char_offset);
1.92 + return char_offset - original_char_offset;
1.93 +}
1.94 +
1.95 +size_t WriteUnicodeCharacter(uint32 code_point, string16* output) {
1.96 + if (CBU16_LENGTH(code_point) == 1) {
1.97 + // Thie code point is in the Basic Multilingual Plane (BMP).
1.98 + output->push_back(static_cast<char16>(code_point));
1.99 + return 1;
1.100 + }
1.101 + // Non-BMP characters use a double-character encoding.
1.102 + size_t char_offset = output->length();
1.103 + output->resize(char_offset + CBU16_MAX_LENGTH);
1.104 + CBU16_APPEND_UNSAFE(&(*output)[0], char_offset, code_point);
1.105 + return CBU16_MAX_LENGTH;
1.106 +}
1.107 +
1.108 +// Generalized Unicode converter -----------------------------------------------
1.109 +
1.110 +template<typename CHAR>
1.111 +void PrepareForUTF8Output(const CHAR* src,
1.112 + size_t src_len,
1.113 + std::string* output) {
1.114 + output->clear();
1.115 + if (src_len == 0)
1.116 + return;
1.117 + if (src[0] < 0x80) {
1.118 + // Assume that the entire input will be ASCII.
1.119 + output->reserve(src_len);
1.120 + } else {
1.121 + // Assume that the entire input is non-ASCII and will have 3 bytes per char.
1.122 + output->reserve(src_len * 3);
1.123 + }
1.124 +}
1.125 +
1.126 +// Instantiate versions we know callers will need.
1.127 +template void PrepareForUTF8Output(const wchar_t*, size_t, std::string*);
1.128 +template void PrepareForUTF8Output(const char16*, size_t, std::string*);
1.129 +
1.130 +template<typename STRING>
1.131 +void PrepareForUTF16Or32Output(const char* src,
1.132 + size_t src_len,
1.133 + STRING* output) {
1.134 + output->clear();
1.135 + if (src_len == 0)
1.136 + return;
1.137 + if (static_cast<unsigned char>(src[0]) < 0x80) {
1.138 + // Assume the input is all ASCII, which means 1:1 correspondence.
1.139 + output->reserve(src_len);
1.140 + } else {
1.141 + // Otherwise assume that the UTF-8 sequences will have 2 bytes for each
1.142 + // character.
1.143 + output->reserve(src_len / 2);
1.144 + }
1.145 +}
1.146 +
1.147 +// Instantiate versions we know callers will need.
1.148 +template void PrepareForUTF16Or32Output(const char*, size_t, std::wstring*);
1.149 +template void PrepareForUTF16Or32Output(const char*, size_t, string16*);
1.150 +
1.151 +} // namespace base
