utf_string_conversions_unittest.cc revision c407dc5cd9bdc5668497f21b26b09d988ab439de
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/basictypes.h" 6#include "base/logging.h" 7#include "base/string_util.h" 8#include "base/utf_string_conversions.h" 9#include "testing/gtest/include/gtest/gtest.h" 10 11namespace base { 12 13namespace { 14 15// Given a null-terminated string of wchar_t with each wchar_t representing 16// a UTF-16 code unit, returns a string16 made up of wchar_t's in the input. 17// Each wchar_t should be <= 0xFFFF and a non-BMP character (> U+FFFF) 18// should be represented as a surrogate pair (two UTF-16 units) 19// *even* where wchar_t is 32-bit (Linux and Mac). 20// 21// This is to help write tests for functions with string16 params until 22// the C++ 0x UTF-16 literal is well-supported by compilers. 23string16 BuildString16(const wchar_t* s) { 24#if defined(WCHAR_T_IS_UTF16) 25 return string16(s); 26#elif defined(WCHAR_T_IS_UTF32) 27 string16 u16; 28 while (*s != 0) { 29 DCHECK_LE(static_cast<unsigned int>(*s), 0xFFFFu); 30 u16.push_back(*s++); 31 } 32 return u16; 33#endif 34} 35 36const wchar_t* const kConvertRoundtripCases[] = { 37 L"Google Video", 38 // "网页 图片 资讯更多 »" 39 L"\x7f51\x9875\x0020\x56fe\x7247\x0020\x8d44\x8baf\x66f4\x591a\x0020\x00bb", 40 // "Παγκόσμιος Ιστός" 41 L"\x03a0\x03b1\x03b3\x03ba\x03cc\x03c3\x03bc\x03b9" 42 L"\x03bf\x03c2\x0020\x0399\x03c3\x03c4\x03cc\x03c2", 43 // "Поиск страниц на русском" 44 L"\x041f\x043e\x0438\x0441\x043a\x0020\x0441\x0442" 45 L"\x0440\x0430\x043d\x0438\x0446\x0020\x043d\x0430" 46 L"\x0020\x0440\x0443\x0441\x0441\x043a\x043e\x043c", 47 // "전체서비스" 48 L"\xc804\xccb4\xc11c\xbe44\xc2a4", 49 50 // Test characters that take more than 16 bits. This will depend on whether 51 // wchar_t is 16 or 32 bits. 52#if defined(WCHAR_T_IS_UTF16) 53 L"\xd800\xdf00", 54 // ????? (Mathematical Alphanumeric Symbols (U+011d40 - U+011d44 : A,B,C,D,E) 55 L"\xd807\xdd40\xd807\xdd41\xd807\xdd42\xd807\xdd43\xd807\xdd44", 56#elif defined(WCHAR_T_IS_UTF32) 57 L"\x10300", 58 // ????? (Mathematical Alphanumeric Symbols (U+011d40 - U+011d44 : A,B,C,D,E) 59 L"\x11d40\x11d41\x11d42\x11d43\x11d44", 60#endif 61}; 62 63} // namespace 64 65TEST(UTFStringConversionsTest, ConvertUTF8AndWide) { 66 // we round-trip all the wide strings through UTF-8 to make sure everything 67 // agrees on the conversion. This uses the stream operators to test them 68 // simultaneously. 69 for (size_t i = 0; i < arraysize(kConvertRoundtripCases); ++i) { 70 std::ostringstream utf8; 71 utf8 << WideToUTF8(kConvertRoundtripCases[i]); 72 std::wostringstream wide; 73 wide << UTF8ToWide(utf8.str()); 74 75 EXPECT_EQ(kConvertRoundtripCases[i], wide.str()); 76 } 77} 78 79TEST(UTFStringConversionsTest, ConvertUTF8AndWideEmptyString) { 80 // An empty std::wstring should be converted to an empty std::string, 81 // and vice versa. 82 std::wstring wempty; 83 std::string empty; 84 EXPECT_EQ(empty, WideToUTF8(wempty)); 85 EXPECT_EQ(wempty, UTF8ToWide(empty)); 86} 87 88TEST(UTFStringConversionsTest, ConvertUTF8ToWide) { 89 struct UTF8ToWideCase { 90 const char* utf8; 91 const wchar_t* wide; 92 bool success; 93 } convert_cases[] = { 94 // Regular UTF-8 input. 95 {"\xe4\xbd\xa0\xe5\xa5\xbd", L"\x4f60\x597d", true}, 96 // Non-character is passed through. 97 {"\xef\xbf\xbfHello", L"\xffffHello", true}, 98 // Truncated UTF-8 sequence. 99 {"\xe4\xa0\xe5\xa5\xbd", L"\xfffd\x597d", false}, 100 // Truncated off the end. 101 {"\xe5\xa5\xbd\xe4\xa0", L"\x597d\xfffd", false}, 102 // Non-shortest-form UTF-8. 103 {"\xf0\x84\xbd\xa0\xe5\xa5\xbd", L"\xfffd\x597d", false}, 104 // This UTF-8 character decodes to a UTF-16 surrogate, which is illegal. 105 {"\xed\xb0\x80", L"\xfffd", false}, 106 // Non-BMP characters. The second is a non-character regarded as valid. 107 // The result will either be in UTF-16 or UTF-32. 108#if defined(WCHAR_T_IS_UTF16) 109 {"A\xF0\x90\x8C\x80z", L"A\xd800\xdf00z", true}, 110 {"A\xF4\x8F\xBF\xBEz", L"A\xdbff\xdffez", true}, 111#elif defined(WCHAR_T_IS_UTF32) 112 {"A\xF0\x90\x8C\x80z", L"A\x10300z", true}, 113 {"A\xF4\x8F\xBF\xBEz", L"A\x10fffez", true}, 114#endif 115 }; 116 117 for (size_t i = 0; i < ARRAYSIZE_UNSAFE(convert_cases); i++) { 118 std::wstring converted; 119 EXPECT_EQ(convert_cases[i].success, 120 UTF8ToWide(convert_cases[i].utf8, 121 strlen(convert_cases[i].utf8), 122 &converted)); 123 std::wstring expected(convert_cases[i].wide); 124 EXPECT_EQ(expected, converted); 125 } 126 127 // Manually test an embedded NULL. 128 std::wstring converted; 129 EXPECT_TRUE(UTF8ToWide("\00Z\t", 3, &converted)); 130 ASSERT_EQ(3U, converted.length()); 131 EXPECT_EQ(static_cast<wchar_t>(0), converted[0]); 132 EXPECT_EQ('Z', converted[1]); 133 EXPECT_EQ('\t', converted[2]); 134 135 // Make sure that conversion replaces, not appends. 136 EXPECT_TRUE(UTF8ToWide("B", 1, &converted)); 137 ASSERT_EQ(1U, converted.length()); 138 EXPECT_EQ('B', converted[0]); 139} 140 141#if defined(WCHAR_T_IS_UTF16) 142// This test is only valid when wchar_t == UTF-16. 143TEST(UTFStringConversionsTest, ConvertUTF16ToUTF8) { 144 struct WideToUTF8Case { 145 const wchar_t* utf16; 146 const char* utf8; 147 bool success; 148 } convert_cases[] = { 149 // Regular UTF-16 input. 150 {L"\x4f60\x597d", "\xe4\xbd\xa0\xe5\xa5\xbd", true}, 151 // Test a non-BMP character. 152 {L"\xd800\xdf00", "\xF0\x90\x8C\x80", true}, 153 // Non-characters are passed through. 154 {L"\xffffHello", "\xEF\xBF\xBFHello", true}, 155 {L"\xdbff\xdffeHello", "\xF4\x8F\xBF\xBEHello", true}, 156 // The first character is a truncated UTF-16 character. 157 {L"\xd800\x597d", "\xef\xbf\xbd\xe5\xa5\xbd", false}, 158 // Truncated at the end. 159 {L"\x597d\xd800", "\xe5\xa5\xbd\xef\xbf\xbd", false}, 160 }; 161 162 for (int i = 0; i < arraysize(convert_cases); i++) { 163 std::string converted; 164 EXPECT_EQ(convert_cases[i].success, 165 WideToUTF8(convert_cases[i].utf16, 166 wcslen(convert_cases[i].utf16), 167 &converted)); 168 std::string expected(convert_cases[i].utf8); 169 EXPECT_EQ(expected, converted); 170 } 171} 172 173#elif defined(WCHAR_T_IS_UTF32) 174// This test is only valid when wchar_t == UTF-32. 175TEST(UTFStringConversionsTest, ConvertUTF32ToUTF8) { 176 struct WideToUTF8Case { 177 const wchar_t* utf32; 178 const char* utf8; 179 bool success; 180 } convert_cases[] = { 181 // Regular 16-bit input. 182 {L"\x4f60\x597d", "\xe4\xbd\xa0\xe5\xa5\xbd", true}, 183 // Test a non-BMP character. 184 {L"A\x10300z", "A\xF0\x90\x8C\x80z", true}, 185 // Non-characters are passed through. 186 {L"\xffffHello", "\xEF\xBF\xBFHello", true}, 187 {L"\x10fffeHello", "\xF4\x8F\xBF\xBEHello", true}, 188 // Invalid Unicode code points. 189 {L"\xfffffffHello", "\xEF\xBF\xBDHello", false}, 190 // The first character is a truncated UTF-16 character. 191 {L"\xd800\x597d", "\xef\xbf\xbd\xe5\xa5\xbd", false}, 192 {L"\xdc01Hello", "\xef\xbf\xbdHello", false}, 193 }; 194 195 for (size_t i = 0; i < ARRAYSIZE_UNSAFE(convert_cases); i++) { 196 std::string converted; 197 EXPECT_EQ(convert_cases[i].success, 198 WideToUTF8(convert_cases[i].utf32, 199 wcslen(convert_cases[i].utf32), 200 &converted)); 201 std::string expected(convert_cases[i].utf8); 202 EXPECT_EQ(expected, converted); 203 } 204} 205#endif // defined(WCHAR_T_IS_UTF32) 206 207TEST(UTFStringConversionsTest, ConvertMultiString) { 208 static wchar_t wmulti[] = { 209 L'f', L'o', L'o', L'\0', 210 L'b', L'a', L'r', L'\0', 211 L'b', L'a', L'z', L'\0', 212 L'\0' 213 }; 214 static char multi[] = { 215 'f', 'o', 'o', '\0', 216 'b', 'a', 'r', '\0', 217 'b', 'a', 'z', '\0', 218 '\0' 219 }; 220 std::wstring wmultistring; 221 memcpy(WriteInto(&wmultistring, arraysize(wmulti)), wmulti, sizeof(wmulti)); 222 EXPECT_EQ(arraysize(wmulti) - 1, wmultistring.length()); 223 std::string expected; 224 memcpy(WriteInto(&expected, arraysize(multi)), multi, sizeof(multi)); 225 EXPECT_EQ(arraysize(multi) - 1, expected.length()); 226 const std::string& converted = WideToUTF8(wmultistring); 227 EXPECT_EQ(arraysize(multi) - 1, converted.length()); 228 EXPECT_EQ(expected, converted); 229} 230 231} // base 232