1 /* 2 * Copyright (C) 2010 The Android Open Source Project 3 * 4 * Licensed under the Apache License, Version 2.0 (the "License"); 5 * you may not use this file except in compliance with the License. 6 * You may obtain a copy of the License at 7 * 8 * http://www.apache.org/licenses/LICENSE-2.0 9 * 10 * Unless required by applicable law or agreed to in writing, software 11 * distributed under the License is distributed on an "AS IS" BASIS, 12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 13 * See the License for the specific language governing permissions and 14 * limitations under the License. 15 */ 16 17 #define LOG_TAG "Unicode_test" 18 19 #include <sys/mman.h> 20 #include <unistd.h> 21 22 #include <log/log.h> 23 #include <utils/Unicode.h> 24 25 #include <gtest/gtest.h> 26 27 namespace android { 28 29 class UnicodeTest : public testing::Test { 30 protected: 31 virtual void SetUp() { 32 } 33 34 virtual void TearDown() { 35 } 36 37 char16_t const * const kSearchString = u"I am a leaf on the wind."; 38 }; 39 40 TEST_F(UnicodeTest, UTF8toUTF16ZeroLength) { 41 ssize_t measured; 42 43 const uint8_t str[] = { }; 44 45 measured = utf8_to_utf16_length(str, 0); 46 EXPECT_EQ(0, measured) 47 << "Zero length input should return zero length output."; 48 } 49 50 TEST_F(UnicodeTest, UTF8toUTF16ASCIILength) { 51 ssize_t measured; 52 53 // U+0030 or ASCII '0' 54 const uint8_t str[] = { 0x30 }; 55 56 measured = utf8_to_utf16_length(str, sizeof(str)); 57 EXPECT_EQ(1, measured) 58 << "ASCII glyphs should have a length of 1 char16_t"; 59 } 60 61 TEST_F(UnicodeTest, UTF8toUTF16Plane1Length) { 62 ssize_t measured; 63 64 // U+2323 SMILE 65 const uint8_t str[] = { 0xE2, 0x8C, 0xA3 }; 66 67 measured = utf8_to_utf16_length(str, sizeof(str)); 68 EXPECT_EQ(1, measured) 69 << "Plane 1 glyphs should have a length of 1 char16_t"; 70 } 71 72 TEST_F(UnicodeTest, UTF8toUTF16SurrogateLength) { 73 ssize_t measured; 74 75 // U+10000 76 const uint8_t str[] = { 0xF0, 0x90, 0x80, 0x80 }; 77 78 measured = utf8_to_utf16_length(str, sizeof(str)); 79 EXPECT_EQ(2, measured) 80 << "Surrogate pairs should have a length of 2 char16_t"; 81 } 82 83 TEST_F(UnicodeTest, UTF8toUTF16TruncatedUTF8) { 84 ssize_t measured; 85 86 // Truncated U+2323 SMILE 87 // U+2323 SMILE 88 const uint8_t str[] = { 0xE2, 0x8C }; 89 90 measured = utf8_to_utf16_length(str, sizeof(str)); 91 EXPECT_EQ(-1, measured) 92 << "Truncated UTF-8 should return -1 to indicate invalid"; 93 } 94 95 TEST_F(UnicodeTest, UTF8toUTF16Normal) { 96 const uint8_t str[] = { 97 0x30, // U+0030, 1 UTF-16 character 98 0xC4, 0x80, // U+0100, 1 UTF-16 character 99 0xE2, 0x8C, 0xA3, // U+2323, 1 UTF-16 character 100 0xF0, 0x90, 0x80, 0x80, // U+10000, 2 UTF-16 character 101 }; 102 103 char16_t output[1 + 1 + 1 + 2 + 1]; // Room for NULL 104 105 utf8_to_utf16(str, sizeof(str), output, sizeof(output) / sizeof(output[0])); 106 107 EXPECT_EQ(0x0030, output[0]) 108 << "should be U+0030"; 109 EXPECT_EQ(0x0100, output[1]) 110 << "should be U+0100"; 111 EXPECT_EQ(0x2323, output[2]) 112 << "should be U+2323"; 113 EXPECT_EQ(0xD800, output[3]) 114 << "should be first half of surrogate U+10000"; 115 EXPECT_EQ(0xDC00, output[4]) 116 << "should be second half of surrogate U+10000"; 117 EXPECT_EQ(NULL, output[5]) 118 << "should be NULL terminated"; 119 } 120 121 TEST_F(UnicodeTest, strstr16EmptyTarget) { 122 EXPECT_EQ(strstr16(kSearchString, u""), kSearchString) 123 << "should return the original pointer"; 124 } 125 126 TEST_F(UnicodeTest, strstr16EmptyTarget_bug) { 127 // In the original code when target is an empty string strlen16() would 128 // start reading the memory until a "terminating null" (that is, zero) 129 // character is found. This happens because "*target++" in the original 130 // code would increment the pointer beyond the actual string. 131 void* memptr; 132 const size_t alignment = sysconf(_SC_PAGESIZE); 133 const size_t size = 2 * alignment; 134 ASSERT_EQ(posix_memalign(&memptr, alignment, size), 0); 135 // Fill allocated memory. 136 memset(memptr, 'A', size); 137 // Create a pointer to an "empty" string on the first page. 138 char16_t* const emptyString = (char16_t* const)((char*)memptr + alignment - 4); 139 *emptyString = (char16_t)0; 140 // Protect the second page to show that strstr16() violates that. 141 ASSERT_EQ(mprotect((char*)memptr + alignment, alignment, PROT_NONE), 0); 142 // Test strstr16(): when bug is present a segmentation fault is raised. 143 ASSERT_EQ(strstr16((char16_t*)memptr, emptyString), (char16_t*)memptr) 144 << "should not read beyond the first char16_t."; 145 // Reset protection of the second page 146 ASSERT_EQ(mprotect((char*)memptr + alignment, alignment, PROT_READ | PROT_WRITE), 0); 147 // Free allocated memory. 148 free(memptr); 149 } 150 151 TEST_F(UnicodeTest, strstr16SameString) { 152 const char16_t* result = strstr16(kSearchString, kSearchString); 153 EXPECT_EQ(kSearchString, result) 154 << "should return the original pointer"; 155 } 156 157 TEST_F(UnicodeTest, strstr16TargetStartOfString) { 158 const char16_t* result = strstr16(kSearchString, u"I am"); 159 EXPECT_EQ(kSearchString, result) 160 << "should return the original pointer"; 161 } 162 163 164 TEST_F(UnicodeTest, strstr16TargetEndOfString) { 165 const char16_t* result = strstr16(kSearchString, u"wind."); 166 EXPECT_EQ(kSearchString+19, result); 167 } 168 169 TEST_F(UnicodeTest, strstr16TargetWithinString) { 170 const char16_t* result = strstr16(kSearchString, u"leaf"); 171 EXPECT_EQ(kSearchString+7, result); 172 } 173 174 TEST_F(UnicodeTest, strstr16TargetNotPresent) { 175 const char16_t* result = strstr16(kSearchString, u"soar"); 176 EXPECT_EQ(nullptr, result); 177 } 178 179 // http://b/29267949 180 // Test that overreading in utf8_to_utf16_length is detected 181 TEST_F(UnicodeTest, InvalidUtf8OverreadDetected) { 182 // An utf8 char starting with \xc4 is two bytes long. 183 // Add extra zeros so no extra memory is read in case the code doesn't 184 // work as expected. 185 static char utf8[] = "\xc4\x00\x00\x00"; 186 ASSERT_DEATH(utf8_to_utf16_length((uint8_t *) utf8, strlen(utf8), 187 true /* overreadIsFatal */), "" /* regex for ASSERT_DEATH */); 188 } 189 190 } 191