base/strings/sys_string_conversions_unittest.cc

   1 // Copyright (c) 2011 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 <string>
   6
   7 #include "base/basictypes.h"
   8 #include "base/strings/string_piece.h"
   9 #include "base/strings/sys_string_conversions.h"
  10 #include "base/strings/utf_string_conversions.h"
  11 #include "base/test/scoped_locale.h"
  12 #include "testing/gtest/include/gtest/gtest.h"
  13
  14 #ifdef WCHAR_T_IS_UTF32
  15 static const std::wstring kSysWideOldItalicLetterA = L"\x10300";
  16 #else
  17 static const std::wstring kSysWideOldItalicLetterA = L"\xd800\xdf00";
  18 #endif
  19
  20 namespace base {
  21
  22 TEST(SysStrings, SysWideToUTF8) {
  23   EXPECT_EQ("Hello, world", SysWideToUTF8(L"Hello, world"));
  24   EXPECT_EQ("\xe4\xbd\xa0\xe5\xa5\xbd", SysWideToUTF8(L"\x4f60\x597d"));
  25
  26   // >16 bits
  27   EXPECT_EQ("\xF0\x90\x8C\x80", SysWideToUTF8(kSysWideOldItalicLetterA));
  28
  29   // Error case. When Windows finds a UTF-16 character going off the end of
  30   // a string, it just converts that literal value to UTF-8, even though this
  31   // is invalid.
  32   //
  33   // This is what XP does, but Vista has different behavior, so we don't bother
  34   // verifying it:
  35   // EXPECT_EQ("\xE4\xBD\xA0\xED\xA0\x80zyxw",
  36   //           SysWideToUTF8(L"\x4f60\xd800zyxw"));
  37
  38   // Test embedded NULLs.
  39   std::wstring wide_null(L"a");
  40   wide_null.push_back(0);
  41   wide_null.push_back('b');
  42
  43   std::string expected_null("a");
  44   expected_null.push_back(0);
  45   expected_null.push_back('b');
  46
  47   EXPECT_EQ(expected_null, SysWideToUTF8(wide_null));
  48 }
  49
  50 TEST(SysStrings, SysUTF8ToWide) {
  51   EXPECT_EQ(L"Hello, world", SysUTF8ToWide("Hello, world"));
  52   EXPECT_EQ(L"\x4f60\x597d", SysUTF8ToWide("\xe4\xbd\xa0\xe5\xa5\xbd"));
  53   // >16 bits
  54   EXPECT_EQ(kSysWideOldItalicLetterA, SysUTF8ToWide("\xF0\x90\x8C\x80"));
  55
  56   // Error case. When Windows finds an invalid UTF-8 character, it just skips
  57   // it. This seems weird because it's inconsistent with the reverse conversion.
  58   //
  59   // This is what XP does, but Vista has different behavior, so we don't bother
  60   // verifying it:
  61   // EXPECT_EQ(L"\x4f60zyxw", SysUTF8ToWide("\xe4\xbd\xa0\xe5\xa5zyxw"));
  62
  63   // Test embedded NULLs.
  64   std::string utf8_null("a");
  65   utf8_null.push_back(0);
  66   utf8_null.push_back('b');
  67
  68   std::wstring expected_null(L"a");
  69   expected_null.push_back(0);
  70   expected_null.push_back('b');
  71
  72   EXPECT_EQ(expected_null, SysUTF8ToWide(utf8_null));
  73 }
  74
  75 #if defined(OS_LINUX)  // Tests depend on setting a specific Linux locale.
  76
  77 TEST(SysStrings, SysWideToNativeMB) {
  78   ScopedLocale locale("en_US.utf-8");
  79   EXPECT_EQ("Hello, world", SysWideToNativeMB(L"Hello, world"));
  80   EXPECT_EQ("\xe4\xbd\xa0\xe5\xa5\xbd", SysWideToNativeMB(L"\x4f60\x597d"));
  81
  82   // >16 bits
  83   EXPECT_EQ("\xF0\x90\x8C\x80", SysWideToNativeMB(kSysWideOldItalicLetterA));
  84
  85   // Error case. When Windows finds a UTF-16 character going off the end of
  86   // a string, it just converts that literal value to UTF-8, even though this
  87   // is invalid.
  88   //
  89   // This is what XP does, but Vista has different behavior, so we don't bother
  90   // verifying it:
  91   // EXPECT_EQ("\xE4\xBD\xA0\xED\xA0\x80zyxw",
  92   //           SysWideToNativeMB(L"\x4f60\xd800zyxw"));
  93
  94   // Test embedded NULLs.
  95   std::wstring wide_null(L"a");
  96   wide_null.push_back(0);
  97   wide_null.push_back('b');
  98
  99   std::string expected_null("a");
 100   expected_null.push_back(0);
 101   expected_null.push_back('b');
 102
 103   EXPECT_EQ(expected_null, SysWideToNativeMB(wide_null));
 104 }
 105
 106 // We assume the test is running in a UTF8 locale.
 107 TEST(SysStrings, SysNativeMBToWide) {
 108   ScopedLocale locale("en_US.utf-8");
 109   EXPECT_EQ(L"Hello, world", SysNativeMBToWide("Hello, world"));
 110   EXPECT_EQ(L"\x4f60\x597d", SysNativeMBToWide("\xe4\xbd\xa0\xe5\xa5\xbd"));
 111   // >16 bits
 112   EXPECT_EQ(kSysWideOldItalicLetterA, SysNativeMBToWide("\xF0\x90\x8C\x80"));
 113
 114   // Error case. When Windows finds an invalid UTF-8 character, it just skips
 115   // it. This seems weird because it's inconsistent with the reverse conversion.
 116   //
 117   // This is what XP does, but Vista has different behavior, so we don't bother
 118   // verifying it:
 119   // EXPECT_EQ(L"\x4f60zyxw", SysNativeMBToWide("\xe4\xbd\xa0\xe5\xa5zyxw"));
 120
 121   // Test embedded NULLs.
 122   std::string utf8_null("a");
 123   utf8_null.push_back(0);
 124   utf8_null.push_back('b');
 125
 126   std::wstring expected_null(L"a");
 127   expected_null.push_back(0);
 128   expected_null.push_back('b');
 129
 130   EXPECT_EQ(expected_null, SysNativeMBToWide(utf8_null));
 131 }
 132
 133 static const wchar_t* const kConvertRoundtripCases[] = {
 134   L"Google Video",
 135   // "网页 图片 资讯更多 »"
 136   L"\x7f51\x9875\x0020\x56fe\x7247\x0020\x8d44\x8baf\x66f4\x591a\x0020\x00bb",
 137   //  "Παγκόσμιος Ιστός"
 138   L"\x03a0\x03b1\x03b3\x03ba\x03cc\x03c3\x03bc\x03b9"
 139   L"\x03bf\x03c2\x0020\x0399\x03c3\x03c4\x03cc\x03c2",
 140   // "Поиск страниц на русском"
 141   L"\x041f\x043e\x0438\x0441\x043a\x0020\x0441\x0442"
 142   L"\x0440\x0430\x043d\x0438\x0446\x0020\x043d\x0430"
 143   L"\x0020\x0440\x0443\x0441\x0441\x043a\x043e\x043c",
 144   // "전체서비스"
 145   L"\xc804\xccb4\xc11c\xbe44\xc2a4",
 146
 147   // Test characters that take more than 16 bits. This will depend on whether
 148   // wchar_t is 16 or 32 bits.
 149 #if defined(WCHAR_T_IS_UTF16)
 150   L"\xd800\xdf00",
 151   // ?????  (Mathematical Alphanumeric Symbols (U+011d40 - U+011d44 : A,B,C,D,E)
 152   L"\xd807\xdd40\xd807\xdd41\xd807\xdd42\xd807\xdd43\xd807\xdd44",
 153 #elif defined(WCHAR_T_IS_UTF32)
 154   L"\x10300",
 155   // ?????  (Mathematical Alphanumeric Symbols (U+011d40 - U+011d44 : A,B,C,D,E)
 156   L"\x11d40\x11d41\x11d42\x11d43\x11d44",
 157 #endif
 158 };
 159
 160
 161 TEST(SysStrings, SysNativeMBAndWide) {
 162   ScopedLocale locale("en_US.utf-8");
 163   for (size_t i = 0; i < arraysize(kConvertRoundtripCases); ++i) {
 164     std::wstring wide = kConvertRoundtripCases[i];
 165     std::wstring trip = SysNativeMBToWide(SysWideToNativeMB(wide));
 166     EXPECT_EQ(wide.size(), trip.size());
 167     EXPECT_EQ(wide, trip);
 168   }
 169
 170   // We assume our test is running in UTF-8, so double check through ICU.
 171   for (size_t i = 0; i < arraysize(kConvertRoundtripCases); ++i) {
 172     std::wstring wide = kConvertRoundtripCases[i];
 173     std::wstring trip = SysNativeMBToWide(WideToUTF8(wide));
 174     EXPECT_EQ(wide.size(), trip.size());
 175     EXPECT_EQ(wide, trip);
 176   }
 177
 178   for (size_t i = 0; i < arraysize(kConvertRoundtripCases); ++i) {
 179     std::wstring wide = kConvertRoundtripCases[i];
 180     std::wstring trip = UTF8ToWide(SysWideToNativeMB(wide));
 181     EXPECT_EQ(wide.size(), trip.size());
 182     EXPECT_EQ(wide, trip);
 183   }
 184 }
 185 #endif  // OS_LINUX
 186
 187 }  // namespace base