• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 // Copyright (c) 2009 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/utf_string_conversion_utils.h"
6 
7 #include "base/third_party/icu/icu_utf.h"
8 #include "util/build_config.h"
9 
10 namespace base {
11 
12 // ReadUnicodeCharacter --------------------------------------------------------
13 
ReadUnicodeCharacter(const char * src,int32_t src_len,int32_t * char_index,uint32_t * code_point_out)14 bool ReadUnicodeCharacter(const char* src,
15                           int32_t src_len,
16                           int32_t* char_index,
17                           uint32_t* code_point_out) {
18   // U8_NEXT expects to be able to use -1 to signal an error, so we must
19   // use a signed type for code_point.  But this function returns false
20   // on error anyway, so code_point_out is unsigned.
21   int32_t code_point;
22   CBU8_NEXT(src, *char_index, src_len, code_point);
23   *code_point_out = static_cast<uint32_t>(code_point);
24 
25   // The ICU macro above moves to the next char, we want to point to the last
26   // char consumed.
27   (*char_index)--;
28 
29   // Validate the decoded value.
30   return IsValidCodepoint(code_point);
31 }
32 
ReadUnicodeCharacter(const char16_t * src,int32_t src_len,int32_t * char_index,uint32_t * code_point)33 bool ReadUnicodeCharacter(const char16_t* src,
34                           int32_t src_len,
35                           int32_t* char_index,
36                           uint32_t* code_point) {
37   if (CBU16_IS_SURROGATE(src[*char_index])) {
38     if (!CBU16_IS_SURROGATE_LEAD(src[*char_index]) ||
39         *char_index + 1 >= src_len || !CBU16_IS_TRAIL(src[*char_index + 1])) {
40       // Invalid surrogate pair.
41       return false;
42     }
43 
44     // Valid surrogate pair.
45     *code_point =
46         CBU16_GET_SUPPLEMENTARY(src[*char_index], src[*char_index + 1]);
47     (*char_index)++;
48   } else {
49     // Not a surrogate, just one 16-bit word.
50     *code_point = src[*char_index];
51   }
52 
53   return IsValidCodepoint(*code_point);
54 }
55 
56 // WriteUnicodeCharacter -------------------------------------------------------
57 
WriteUnicodeCharacter(uint32_t code_point,std::string * output)58 size_t WriteUnicodeCharacter(uint32_t code_point, std::string* output) {
59   if (code_point <= 0x7f) {
60     // Fast path the common case of one byte.
61     output->push_back(static_cast<char>(code_point));
62     return 1;
63   }
64 
65   // CBU8_APPEND_UNSAFE can append up to 4 bytes.
66   size_t char_offset = output->length();
67   size_t original_char_offset = char_offset;
68   output->resize(char_offset + CBU8_MAX_LENGTH);
69 
70   CBU8_APPEND_UNSAFE(&(*output)[0], char_offset, code_point);
71 
72   // CBU8_APPEND_UNSAFE will advance our pointer past the inserted character, so
73   // it will represent the new length of the string.
74   output->resize(char_offset);
75   return char_offset - original_char_offset;
76 }
77 
WriteUnicodeCharacter(uint32_t code_point,std::u16string * output)78 size_t WriteUnicodeCharacter(uint32_t code_point, std::u16string* output) {
79   if (CBU16_LENGTH(code_point) == 1) {
80     // Thie code point is in the Basic Multilingual Plane (BMP).
81     output->push_back(static_cast<char16_t>(code_point));
82     return 1;
83   }
84   // Non-BMP characters use a double-character encoding.
85   size_t char_offset = output->length();
86   output->resize(char_offset + CBU16_MAX_LENGTH);
87   CBU16_APPEND_UNSAFE(&(*output)[0], char_offset, code_point);
88   return CBU16_MAX_LENGTH;
89 }
90 
91 // Generalized Unicode converter -----------------------------------------------
92 
93 template <typename CHAR>
PrepareForUTF8Output(const CHAR * src,size_t src_len,std::string * output)94 void PrepareForUTF8Output(const CHAR* src,
95                           size_t src_len,
96                           std::string* output) {
97   output->clear();
98   if (src_len == 0)
99     return;
100   if (src[0] < 0x80) {
101     // Assume that the entire input will be ASCII.
102     output->reserve(src_len);
103   } else {
104     // Assume that the entire input is non-ASCII and will have 3 bytes per char.
105     output->reserve(src_len * 3);
106   }
107 }
108 
109 // Instantiate versions we know callers will need.
110 template void PrepareForUTF8Output(const char16_t*, size_t, std::string*);
111 
112 template <typename STRING>
PrepareForUTF16Or32Output(const char * src,size_t src_len,STRING * output)113 void PrepareForUTF16Or32Output(const char* src,
114                                size_t src_len,
115                                STRING* output) {
116   output->clear();
117   if (src_len == 0)
118     return;
119   if (static_cast<unsigned char>(src[0]) < 0x80) {
120     // Assume the input is all ASCII, which means 1:1 correspondence.
121     output->reserve(src_len);
122   } else {
123     // Otherwise assume that the UTF-8 sequences will have 2 bytes for each
124     // character.
125     output->reserve(src_len / 2);
126   }
127 }
128 
129 // Instantiate versions we know callers will need.
130 template void PrepareForUTF16Or32Output(const char*, size_t, std::u16string*);
131 
132 }  // namespace base
133