1 // Copyright 2010 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 "base/strings/utf_string_conversions.h"
6
7 #include <stddef.h>
8
9 #include "base/strings/string_piece.h"
10 #include "base/strings/string_util.h"
11 #include "build/build_config.h"
12 #include "testing/gtest/include/gtest/gtest.h"
13
14 namespace base {
15
16 namespace {
17
18 const wchar_t* const kConvertRoundtripCases[] = {
19 L"Google Video",
20 // "网页 图片 资讯更多 »"
21 L"\x7f51\x9875\x0020\x56fe\x7247\x0020\x8d44\x8baf\x66f4\x591a\x0020\x00bb",
22 // "Παγκόσμιος Ιστός"
23 L"\x03a0\x03b1\x03b3\x03ba\x03cc\x03c3\x03bc\x03b9"
24 L"\x03bf\x03c2\x0020\x0399\x03c3\x03c4\x03cc\x03c2",
25 // "Поиск страниц на русском"
26 L"\x041f\x043e\x0438\x0441\x043a\x0020\x0441\x0442"
27 L"\x0440\x0430\x043d\x0438\x0446\x0020\x043d\x0430"
28 L"\x0020\x0440\x0443\x0441\x0441\x043a\x043e\x043c",
29 // "전체서비스"
30 L"\xc804\xccb4\xc11c\xbe44\xc2a4",
31
32 // Test characters that take more than 16 bits. This will depend on whether
33 // wchar_t is 16 or 32 bits.
34 #if defined(WCHAR_T_IS_16_BIT)
35 L"\xd800\xdf00",
36 // ????? (Mathematical Alphanumeric Symbols (U+011d40 - U+011d44 :
37 // A,B,C,D,E)
38 L"\xd807\xdd40\xd807\xdd41\xd807\xdd42\xd807\xdd43\xd807\xdd44",
39 #elif defined(WCHAR_T_IS_32_BIT)
40 L"\x10300",
41 // ????? (Mathematical Alphanumeric Symbols (U+011d40 - U+011d44 :
42 // A,B,C,D,E)
43 L"\x11d40\x11d41\x11d42\x11d43\x11d44",
44 #endif
45 };
46
47 } // namespace
48
TEST(UTFStringConversionsTest,ConvertUTF8AndWide)49 TEST(UTFStringConversionsTest, ConvertUTF8AndWide) {
50 // we round-trip all the wide strings through UTF-8 to make sure everything
51 // agrees on the conversion. This uses the stream operators to test them
52 // simultaneously.
53 for (auto* i : kConvertRoundtripCases) {
54 std::ostringstream utf8;
55 utf8 << WideToUTF8(i);
56 std::wostringstream wide;
57 wide << UTF8ToWide(utf8.str());
58
59 EXPECT_EQ(i, wide.str());
60 }
61 }
62
TEST(UTFStringConversionsTest,ConvertUTF8AndWideEmptyString)63 TEST(UTFStringConversionsTest, ConvertUTF8AndWideEmptyString) {
64 // An empty std::wstring should be converted to an empty std::string,
65 // and vice versa.
66 std::wstring wempty;
67 std::string empty;
68 EXPECT_EQ(empty, WideToUTF8(wempty));
69 EXPECT_EQ(wempty, UTF8ToWide(empty));
70 }
71
TEST(UTFStringConversionsTest,ConvertUTF8ToWide)72 TEST(UTFStringConversionsTest, ConvertUTF8ToWide) {
73 struct UTF8ToWideCase {
74 const char* utf8;
75 const wchar_t* wide;
76 bool success;
77 } convert_cases[] = {
78 // Regular UTF-8 input.
79 {"\xe4\xbd\xa0\xe5\xa5\xbd", L"\x4f60\x597d", true},
80 // Non-character is passed through.
81 {"\xef\xbf\xbfHello", L"\xffffHello", true},
82 // Truncated UTF-8 sequence.
83 {"\xe4\xa0\xe5\xa5\xbd", L"\xfffd\x597d", false},
84 // Truncated off the end.
85 {"\xe5\xa5\xbd\xe4\xa0", L"\x597d\xfffd", false},
86 // Non-shortest-form UTF-8.
87 {"\xf0\x84\xbd\xa0\xe5\xa5\xbd", L"\xfffd\xfffd\xfffd\xfffd\x597d", false},
88 // This UTF-8 character decodes to a UTF-16 surrogate, which is illegal.
89 {"\xed\xb0\x80", L"\xfffd\xfffd\xfffd", false},
90 // Non-BMP characters. The second is a non-character regarded as valid.
91 // The result will either be in UTF-16 or UTF-32.
92 #if defined(WCHAR_T_IS_16_BIT)
93 {"A\xF0\x90\x8C\x80z", L"A\xd800\xdf00z", true},
94 {"A\xF4\x8F\xBF\xBEz", L"A\xdbff\xdffez", true},
95 #elif defined(WCHAR_T_IS_32_BIT)
96 {"A\xF0\x90\x8C\x80z", L"A\x10300z", true},
97 {"A\xF4\x8F\xBF\xBEz", L"A\x10fffez", true},
98 #endif
99 };
100
101 for (const auto& i : convert_cases) {
102 std::wstring converted;
103 EXPECT_EQ(i.success, UTF8ToWide(i.utf8, strlen(i.utf8), &converted));
104 std::wstring expected(i.wide);
105 EXPECT_EQ(expected, converted);
106 }
107
108 // Manually test an embedded NULL.
109 std::wstring converted;
110 EXPECT_TRUE(UTF8ToWide("\00Z\t", 3, &converted));
111 ASSERT_EQ(3U, converted.length());
112 EXPECT_EQ(static_cast<wchar_t>(0), converted[0]);
113 EXPECT_EQ('Z', converted[1]);
114 EXPECT_EQ('\t', converted[2]);
115
116 // Make sure that conversion replaces, not appends.
117 EXPECT_TRUE(UTF8ToWide("B", 1, &converted));
118 ASSERT_EQ(1U, converted.length());
119 EXPECT_EQ('B', converted[0]);
120 }
121
122 #if defined(WCHAR_T_IS_16_BIT)
123 // This test is only valid when wchar_t == UTF-16.
TEST(UTFStringConversionsTest,ConvertUTF16ToUTF8)124 TEST(UTFStringConversionsTest, ConvertUTF16ToUTF8) {
125 struct WideToUTF8Case {
126 const wchar_t* utf16;
127 const char* utf8;
128 bool success;
129 } convert_cases[] = {
130 // Regular UTF-16 input.
131 {L"\x4f60\x597d", "\xe4\xbd\xa0\xe5\xa5\xbd", true},
132 // Test a non-BMP character.
133 {L"\xd800\xdf00", "\xF0\x90\x8C\x80", true},
134 // Non-characters are passed through.
135 {L"\xffffHello", "\xEF\xBF\xBFHello", true},
136 {L"\xdbff\xdffeHello", "\xF4\x8F\xBF\xBEHello", true},
137 // The first character is a truncated UTF-16 character.
138 {L"\xd800\x597d", "\xef\xbf\xbd\xe5\xa5\xbd", false},
139 // Truncated at the end.
140 {L"\x597d\xd800", "\xe5\xa5\xbd\xef\xbf\xbd", false},
141 };
142
143 for (const auto& test : convert_cases) {
144 std::string converted;
145 EXPECT_EQ(test.success,
146 WideToUTF8(test.utf16, wcslen(test.utf16), &converted));
147 std::string expected(test.utf8);
148 EXPECT_EQ(expected, converted);
149 }
150 }
151
152 #elif defined(WCHAR_T_IS_32_BIT)
153 // This test is only valid when wchar_t == UTF-32.
TEST(UTFStringConversionsTest,ConvertUTF32ToUTF8)154 TEST(UTFStringConversionsTest, ConvertUTF32ToUTF8) {
155 struct WideToUTF8Case {
156 const wchar_t* utf32;
157 const char* utf8;
158 bool success;
159 } convert_cases[] = {
160 // Regular 16-bit input.
161 {L"\x4f60\x597d", "\xe4\xbd\xa0\xe5\xa5\xbd", true},
162 // Test a non-BMP character.
163 {L"A\x10300z", "A\xF0\x90\x8C\x80z", true},
164 // Non-characters are passed through.
165 {L"\xffffHello", "\xEF\xBF\xBFHello", true},
166 {L"\x10fffeHello", "\xF4\x8F\xBF\xBEHello", true},
167 // Invalid Unicode code points.
168 {L"\xfffffffHello", "\xEF\xBF\xBDHello", false},
169 // The first character is a truncated UTF-16 character.
170 {L"\xd800\x597d", "\xef\xbf\xbd\xe5\xa5\xbd", false},
171 {L"\xdc01Hello", "\xef\xbf\xbdHello", false},
172 };
173
174 for (const auto& test : convert_cases) {
175 std::string converted;
176 EXPECT_EQ(test.success,
177 WideToUTF8(test.utf32, wcslen(test.utf32), &converted));
178 std::string expected(test.utf8);
179 EXPECT_EQ(expected, converted);
180 }
181 }
182 #endif // defined(WCHAR_T_IS_32_BIT)
183
TEST(UTFStringConversionsTest,ConvertMultiString)184 TEST(UTFStringConversionsTest, ConvertMultiString) {
185 static char16_t multi16[] = {'f', 'o', 'o', '\0', 'b', 'a', 'r',
186 '\0', 'b', 'a', 'z', '\0', '\0'};
187 static char multi[] = {
188 'f', 'o', 'o', '\0',
189 'b', 'a', 'r', '\0',
190 'b', 'a', 'z', '\0',
191 '\0'
192 };
193 std::u16string multistring16;
194 memcpy(WriteInto(&multistring16, std::size(multi16)), multi16,
195 sizeof(multi16));
196 EXPECT_EQ(std::size(multi16) - 1, multistring16.length());
197 std::string expected;
198 memcpy(WriteInto(&expected, std::size(multi)), multi, sizeof(multi));
199 EXPECT_EQ(std::size(multi) - 1, expected.length());
200 const std::string& converted = UTF16ToUTF8(multistring16);
201 EXPECT_EQ(std::size(multi) - 1, converted.length());
202 EXPECT_EQ(expected, converted);
203 }
204
205 } // namespace base
206