• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 // Copyright 2011 The Chromium Authors
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 <stddef.h>
6 
7 #include <string>
8 
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 "build/build_config.h"
13 #include "testing/gtest/include/gtest/gtest.h"
14 
15 #ifdef WCHAR_T_IS_32_BIT
16 static const std::wstring kSysWideOldItalicLetterA = L"\x10300";
17 #else
18 static const std::wstring kSysWideOldItalicLetterA = L"\xd800\xdf00";
19 #endif
20 
21 namespace base {
22 
TEST(SysStrings,SysWideToUTF8)23 TEST(SysStrings, SysWideToUTF8) {
24   EXPECT_EQ("Hello, world", SysWideToUTF8(L"Hello, world"));
25   EXPECT_EQ("\xe4\xbd\xa0\xe5\xa5\xbd", SysWideToUTF8(L"\x4f60\x597d"));
26 
27   // A value outside of the BMP and therefore not representable with one UTF-16
28   // code unit.
29   EXPECT_EQ("\xF0\x90\x8C\x80", SysWideToUTF8(kSysWideOldItalicLetterA));
30 
31   // Error case. When Windows finds a UTF-16 character going off the end of
32   // a string, it just converts that literal value to UTF-8, even though this
33   // is invalid.
34   //
35   // This is what XP does, but Vista has different behavior, so we don't bother
36   // verifying it:
37   // EXPECT_EQ("\xE4\xBD\xA0\xED\xA0\x80zyxw",
38   //           SysWideToUTF8(L"\x4f60\xd800zyxw"));
39 
40   // Test embedded NULLs.
41   std::wstring wide_null(L"a");
42   wide_null.push_back(0);
43   wide_null.push_back('b');
44 
45   std::string expected_null("a");
46   expected_null.push_back(0);
47   expected_null.push_back('b');
48 
49   EXPECT_EQ(expected_null, SysWideToUTF8(wide_null));
50 }
51 
TEST(SysStrings,SysUTF8ToWide)52 TEST(SysStrings, SysUTF8ToWide) {
53   EXPECT_EQ(L"Hello, world", SysUTF8ToWide("Hello, world"));
54   EXPECT_EQ(L"\x4f60\x597d", SysUTF8ToWide("\xe4\xbd\xa0\xe5\xa5\xbd"));
55 
56   // A value outside of the BMP and therefore not representable with one UTF-16
57   // code unit.
58   EXPECT_EQ(kSysWideOldItalicLetterA, SysUTF8ToWide("\xF0\x90\x8C\x80"));
59 
60   // Error case. When Windows finds an invalid UTF-8 character, it just skips
61   // it. This seems weird because it's inconsistent with the reverse conversion.
62   //
63   // This is what XP does, but Vista has different behavior, so we don't bother
64   // verifying it:
65   // EXPECT_EQ(L"\x4f60zyxw", SysUTF8ToWide("\xe4\xbd\xa0\xe5\xa5zyxw"));
66 
67   // Test embedded NULLs.
68   std::string utf8_null("a");
69   utf8_null.push_back(0);
70   utf8_null.push_back('b');
71 
72   std::wstring expected_null(L"a");
73   expected_null.push_back(0);
74   expected_null.push_back('b');
75 
76   EXPECT_EQ(expected_null, SysUTF8ToWide(utf8_null));
77 }
78 
79 // Tests depend on setting a specific Linux locale.
80 #if BUILDFLAG(IS_LINUX) || BUILDFLAG(IS_CHROMEOS)
TEST(SysStrings,SysWideToNativeMB)81 TEST(SysStrings, SysWideToNativeMB) {
82 #if !defined(SYSTEM_NATIVE_UTF8)
83   ScopedLocale locale("en_US.UTF-8");
84 #endif
85   EXPECT_EQ("Hello, world", SysWideToNativeMB(L"Hello, world"));
86   EXPECT_EQ("\xe4\xbd\xa0\xe5\xa5\xbd", SysWideToNativeMB(L"\x4f60\x597d"));
87 
88   // A value outside of the BMP and therefore not representable with one UTF-16
89   // code unit.
90   EXPECT_EQ("\xF0\x90\x8C\x80", SysWideToNativeMB(kSysWideOldItalicLetterA));
91 
92   // Error case. When Windows finds a UTF-16 character going off the end of
93   // a string, it just converts that literal value to UTF-8, even though this
94   // is invalid.
95   //
96   // This is what XP does, but Vista has different behavior, so we don't bother
97   // verifying it:
98   // EXPECT_EQ("\xE4\xBD\xA0\xED\xA0\x80zyxw",
99   //           SysWideToNativeMB(L"\x4f60\xd800zyxw"));
100 
101   // Test embedded NULLs.
102   std::wstring wide_null(L"a");
103   wide_null.push_back(0);
104   wide_null.push_back('b');
105 
106   std::string expected_null("a");
107   expected_null.push_back(0);
108   expected_null.push_back('b');
109 
110   EXPECT_EQ(expected_null, SysWideToNativeMB(wide_null));
111 }
112 
113 // We assume the test is running in a UTF8 locale.
TEST(SysStrings,SysNativeMBToWide)114 TEST(SysStrings, SysNativeMBToWide) {
115 #if !defined(SYSTEM_NATIVE_UTF8)
116   ScopedLocale locale("en_US.UTF-8");
117 #endif
118   EXPECT_EQ(L"Hello, world", SysNativeMBToWide("Hello, world"));
119   EXPECT_EQ(L"\x4f60\x597d", SysNativeMBToWide("\xe4\xbd\xa0\xe5\xa5\xbd"));
120 
121   // A value outside of the BMP and therefore not representable with one UTF-16
122   // code unit.
123   EXPECT_EQ(kSysWideOldItalicLetterA, SysNativeMBToWide("\xF0\x90\x8C\x80"));
124 
125   // Error case. When Windows finds an invalid UTF-8 character, it just skips
126   // it. This seems weird because it's inconsistent with the reverse conversion.
127   //
128   // This is what XP does, but Vista has different behavior, so we don't bother
129   // verifying it:
130   // EXPECT_EQ(L"\x4f60zyxw", SysNativeMBToWide("\xe4\xbd\xa0\xe5\xa5zyxw"));
131 
132   // Test embedded NULLs.
133   std::string utf8_null("a");
134   utf8_null.push_back(0);
135   utf8_null.push_back('b');
136 
137   std::wstring expected_null(L"a");
138   expected_null.push_back(0);
139   expected_null.push_back('b');
140 
141   EXPECT_EQ(expected_null, SysNativeMBToWide(utf8_null));
142 }
143 
144 static const wchar_t* const kConvertRoundtripCases[] = {
145     L"Google Video",
146     // "网页 图片 资讯更多 »"
147     L"\x7f51\x9875\x0020\x56fe\x7247\x0020\x8d44\x8baf\x66f4\x591a\x0020\x00bb",
148     //  "Παγκόσμιος Ιστός"
149     L"\x03a0\x03b1\x03b3\x03ba\x03cc\x03c3\x03bc\x03b9"
150     L"\x03bf\x03c2\x0020\x0399\x03c3\x03c4\x03cc\x03c2",
151     // "Поиск страниц на русском"
152     L"\x041f\x043e\x0438\x0441\x043a\x0020\x0441\x0442"
153     L"\x0440\x0430\x043d\x0438\x0446\x0020\x043d\x0430"
154     L"\x0020\x0440\x0443\x0441\x0441\x043a\x043e\x043c",
155     // "전체서비스"
156     L"\xc804\xccb4\xc11c\xbe44\xc2a4",
157 
158 // Test characters that take more than 16 bits. This will depend on whether
159 // wchar_t is 16 or 32 bits.
160 #if defined(WCHAR_T_IS_16_BIT)
161     L"\xd800\xdf00",
162     // ?????  (Mathematical Alphanumeric Symbols (U+011d40 - U+011d44 :
163     // A,B,C,D,E)
164     L"\xd807\xdd40\xd807\xdd41\xd807\xdd42\xd807\xdd43\xd807\xdd44",
165 #elif defined(WCHAR_T_IS_32_BIT)
166     L"\x10300",
167     // ?????  (Mathematical Alphanumeric Symbols (U+011d40 - U+011d44 :
168     // A,B,C,D,E)
169     L"\x11d40\x11d41\x11d42\x11d43\x11d44",
170 #endif
171 };
172 
173 
TEST(SysStrings,SysNativeMBAndWide)174 TEST(SysStrings, SysNativeMBAndWide) {
175 #if !defined(SYSTEM_NATIVE_UTF8)
176   ScopedLocale locale("en_US.UTF-8");
177 #endif
178   for (auto* i : kConvertRoundtripCases) {
179     std::wstring wide = i;
180     std::wstring trip = SysNativeMBToWide(SysWideToNativeMB(wide));
181     EXPECT_EQ(wide.size(), trip.size());
182     EXPECT_EQ(wide, trip);
183   }
184 
185   // We assume our test is running in UTF-8, so double check through ICU.
186   for (auto* i : kConvertRoundtripCases) {
187     std::wstring wide = i;
188     std::wstring trip = SysNativeMBToWide(WideToUTF8(wide));
189     EXPECT_EQ(wide.size(), trip.size());
190     EXPECT_EQ(wide, trip);
191   }
192 
193   for (auto* i : kConvertRoundtripCases) {
194     std::wstring wide = i;
195     std::wstring trip = UTF8ToWide(SysWideToNativeMB(wide));
196     EXPECT_EQ(wide.size(), trip.size());
197     EXPECT_EQ(wide, trip);
198   }
199 }
200 #endif  // BUILDFLAG(IS_LINUX) || BUILDFLAG(IS_CHROMEOS)
201 
202 }  // namespace base
203