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 <algorithm> 6 7 #include "base/logging.h" 8 #include "base/strings/string_piece.h" 9 #include "base/strings/utf_offset_string_conversions.h" 10 #include "testing/gtest/include/gtest/gtest.h" 11 12 namespace base { 13 14 namespace { 15 16 static const size_t kNpos = string16::npos; 17 18 } // namespace 19 20 TEST(UTFOffsetStringConversionsTest, AdjustOffset) { 21 struct UTF8ToUTF16Case { 22 const char* utf8; 23 size_t input_offset; 24 size_t output_offset; 25 } utf8_to_utf16_cases[] = { 26 {"", 0, 0}, 27 {"", kNpos, kNpos}, 28 {"\xe4\xbd\xa0\xe5\xa5\xbd", 1, kNpos}, 29 {"\xe4\xbd\xa0\xe5\xa5\xbd", 3, 1}, 30 {"\xed\xb0\x80z", 3, 1}, 31 {"A\xF0\x90\x8C\x80z", 1, 1}, 32 {"A\xF0\x90\x8C\x80z", 2, kNpos}, 33 {"A\xF0\x90\x8C\x80z", 5, 3}, 34 {"A\xF0\x90\x8C\x80z", 6, 4}, 35 {"A\xF0\x90\x8C\x80z", kNpos, kNpos}, 36 }; 37 for (size_t i = 0; i < ARRAYSIZE_UNSAFE(utf8_to_utf16_cases); ++i) { 38 const size_t offset = utf8_to_utf16_cases[i].input_offset; 39 std::vector<size_t> offsets; 40 offsets.push_back(offset); 41 UTF8ToUTF16AndAdjustOffsets(utf8_to_utf16_cases[i].utf8, &offsets); 42 EXPECT_EQ(utf8_to_utf16_cases[i].output_offset, offsets[0]); 43 } 44 45 struct UTF16ToUTF8Case { 46 char16 utf16[10]; 47 size_t input_offset; 48 size_t output_offset; 49 } utf16_to_utf8_cases[] = { 50 {{}, 0, 0}, 51 // Converted to 3-byte utf-8 sequences 52 {{0x5909, 0x63DB}, 3, kNpos}, 53 {{0x5909, 0x63DB}, 2, 6}, 54 {{0x5909, 0x63DB}, 1, 3}, 55 {{0x5909, 0x63DB}, 0, 0}, 56 // Converted to 2-byte utf-8 sequences 57 {{'A', 0x00bc, 0x00be, 'z'}, 1, 1}, 58 {{'A', 0x00bc, 0x00be, 'z'}, 2, 3}, 59 {{'A', 0x00bc, 0x00be, 'z'}, 3, 5}, 60 {{'A', 0x00bc, 0x00be, 'z'}, 4, 6}, 61 // Surrogate pair 62 {{'A', 0xd800, 0xdf00, 'z'}, 1, 1}, 63 {{'A', 0xd800, 0xdf00, 'z'}, 2, kNpos}, 64 {{'A', 0xd800, 0xdf00, 'z'}, 3, 5}, 65 {{'A', 0xd800, 0xdf00, 'z'}, 4, 6}, 66 }; 67 for (size_t i = 0; i < ARRAYSIZE_UNSAFE(utf16_to_utf8_cases); ++i) { 68 size_t offset = utf16_to_utf8_cases[i].input_offset; 69 std::vector<size_t> offsets; 70 offsets.push_back(offset); 71 UTF16ToUTF8AndAdjustOffsets(utf16_to_utf8_cases[i].utf16, &offsets); 72 EXPECT_EQ(utf16_to_utf8_cases[i].output_offset, offsets[0]) << i; 73 } 74 } 75 76 TEST(UTFOffsetStringConversionsTest, LimitOffsets) { 77 const size_t kLimit = 10; 78 const size_t kItems = 20; 79 std::vector<size_t> size_ts; 80 for (size_t t = 0; t < kItems; ++t) 81 size_ts.push_back(t); 82 std::for_each(size_ts.begin(), size_ts.end(), 83 LimitOffset<string16>(kLimit)); 84 size_t unlimited_count = 0; 85 for (std::vector<size_t>::iterator ti = size_ts.begin(); ti != size_ts.end(); 86 ++ti) { 87 if (*ti != kNpos) 88 ++unlimited_count; 89 } 90 EXPECT_EQ(11U, unlimited_count); 91 92 // Reverse the values in the vector and try again. 93 size_ts.clear(); 94 for (size_t t = kItems; t > 0; --t) 95 size_ts.push_back(t - 1); 96 std::for_each(size_ts.begin(), size_ts.end(), 97 LimitOffset<string16>(kLimit)); 98 unlimited_count = 0; 99 for (std::vector<size_t>::iterator ti = size_ts.begin(); ti != size_ts.end(); 100 ++ti) { 101 if (*ti != kNpos) 102 ++unlimited_count; 103 } 104 EXPECT_EQ(11U, unlimited_count); 105 } 106 107 TEST(UTFOffsetStringConversionsTest, AdjustOffsets) { 108 // Imagine we have strings as shown in the following cases where the 109 // X's represent encoded characters. 110 // 1: abcXXXdef ==> abcXdef 111 { 112 std::vector<size_t> offsets; 113 for (size_t t = 0; t <= 9; ++t) 114 offsets.push_back(t); 115 OffsetAdjuster::Adjustments adjustments; 116 adjustments.push_back(OffsetAdjuster::Adjustment(3, 3, 1)); 117 OffsetAdjuster::AdjustOffsets(adjustments, &offsets); 118 size_t expected_1[] = {0, 1, 2, 3, kNpos, kNpos, 4, 5, 6, 7}; 119 EXPECT_EQ(offsets.size(), arraysize(expected_1)); 120 for (size_t i = 0; i < arraysize(expected_1); ++i) 121 EXPECT_EQ(expected_1[i], offsets[i]); 122 } 123 124 // 2: XXXaXXXXbcXXXXXXXdefXXX ==> XaXXbcXXXXdefX 125 { 126 std::vector<size_t> offsets; 127 for (size_t t = 0; t <= 23; ++t) 128 offsets.push_back(t); 129 OffsetAdjuster::Adjustments adjustments; 130 adjustments.push_back(OffsetAdjuster::Adjustment(0, 3, 1)); 131 adjustments.push_back(OffsetAdjuster::Adjustment(4, 4, 2)); 132 adjustments.push_back(OffsetAdjuster::Adjustment(10, 7, 4)); 133 adjustments.push_back(OffsetAdjuster::Adjustment(20, 3, 1)); 134 OffsetAdjuster::AdjustOffsets(adjustments, &offsets); 135 size_t expected_2[] = { 136 0, kNpos, kNpos, 1, 2, kNpos, kNpos, kNpos, 4, 5, 6, kNpos, kNpos, kNpos, 137 kNpos, kNpos, kNpos, 10, 11, 12, 13, kNpos, kNpos, 14 138 }; 139 EXPECT_EQ(offsets.size(), arraysize(expected_2)); 140 for (size_t i = 0; i < arraysize(expected_2); ++i) 141 EXPECT_EQ(expected_2[i], offsets[i]); 142 } 143 144 // 3: XXXaXXXXbcdXXXeXX ==> aXXXXbcdXXXe 145 { 146 std::vector<size_t> offsets; 147 for (size_t t = 0; t <= 17; ++t) 148 offsets.push_back(t); 149 OffsetAdjuster::Adjustments adjustments; 150 adjustments.push_back(OffsetAdjuster::Adjustment(0, 3, 0)); 151 adjustments.push_back(OffsetAdjuster::Adjustment(4, 4, 4)); 152 adjustments.push_back(OffsetAdjuster::Adjustment(11, 3, 3)); 153 adjustments.push_back(OffsetAdjuster::Adjustment(15, 2, 0)); 154 OffsetAdjuster::AdjustOffsets(adjustments, &offsets); 155 size_t expected_3[] = { 156 0, kNpos, kNpos, 0, 1, kNpos, kNpos, kNpos, 5, 6, 7, 8, kNpos, kNpos, 11, 157 12, kNpos, 12 158 }; 159 EXPECT_EQ(offsets.size(), arraysize(expected_3)); 160 for (size_t i = 0; i < arraysize(expected_3); ++i) 161 EXPECT_EQ(expected_3[i], offsets[i]); 162 } 163 } 164 165 TEST(UTFOffsetStringConversionsTest, UnadjustOffsets) { 166 // Imagine we have strings as shown in the following cases where the 167 // X's represent encoded characters. 168 // 1: abcXXXdef ==> abcXdef 169 { 170 std::vector<size_t> offsets; 171 for (size_t t = 0; t <= 7; ++t) 172 offsets.push_back(t); 173 OffsetAdjuster::Adjustments adjustments; 174 adjustments.push_back(OffsetAdjuster::Adjustment(3, 3, 1)); 175 OffsetAdjuster::UnadjustOffsets(adjustments, &offsets); 176 size_t expected_1[] = {0, 1, 2, 3, 6, 7, 8, 9}; 177 EXPECT_EQ(offsets.size(), arraysize(expected_1)); 178 for (size_t i = 0; i < arraysize(expected_1); ++i) 179 EXPECT_EQ(expected_1[i], offsets[i]); 180 } 181 182 // 2: XXXaXXXXbcXXXXXXXdefXXX ==> XaXXbcXXXXdefX 183 { 184 std::vector<size_t> offsets; 185 for (size_t t = 0; t <= 14; ++t) 186 offsets.push_back(t); 187 OffsetAdjuster::Adjustments adjustments; 188 adjustments.push_back(OffsetAdjuster::Adjustment(0, 3, 1)); 189 adjustments.push_back(OffsetAdjuster::Adjustment(4, 4, 2)); 190 adjustments.push_back(OffsetAdjuster::Adjustment(10, 7, 4)); 191 adjustments.push_back(OffsetAdjuster::Adjustment(20, 3, 1)); 192 OffsetAdjuster::UnadjustOffsets(adjustments, &offsets); 193 size_t expected_2[] = { 194 0, 3, 4, kNpos, 8, 9, 10, kNpos, kNpos, kNpos, 17, 18, 19, 20, 23 195 }; 196 EXPECT_EQ(offsets.size(), arraysize(expected_2)); 197 for (size_t i = 0; i < arraysize(expected_2); ++i) 198 EXPECT_EQ(expected_2[i], offsets[i]); 199 } 200 201 // 3: XXXaXXXXbcdXXXeXX ==> aXXXXbcdXXXe 202 { 203 std::vector<size_t> offsets; 204 for (size_t t = 0; t <= 12; ++t) 205 offsets.push_back(t); 206 OffsetAdjuster::Adjustments adjustments; 207 adjustments.push_back(OffsetAdjuster::Adjustment(0, 3, 0)); 208 adjustments.push_back(OffsetAdjuster::Adjustment(4, 4, 4)); 209 adjustments.push_back(OffsetAdjuster::Adjustment(11, 3, 3)); 210 adjustments.push_back(OffsetAdjuster::Adjustment(15, 2, 0)); 211 OffsetAdjuster::UnadjustOffsets(adjustments, &offsets); 212 size_t expected_3[] = { 213 0, // this could just as easily be 3 214 4, kNpos, kNpos, kNpos, 8, 9, 10, 11, kNpos, kNpos, 14, 215 15 // this could just as easily be 17 216 }; 217 EXPECT_EQ(offsets.size(), arraysize(expected_3)); 218 for (size_t i = 0; i < arraysize(expected_3); ++i) 219 EXPECT_EQ(expected_3[i], offsets[i]); 220 } 221 } 222 223 // MergeSequentialAdjustments is used by net/base/escape.{h,cc} and 224 // net/base/net_util.{h,cc}. The two tests EscapeTest.AdjustOffset and 225 // NetUtilTest.FormatUrlWithOffsets test its behavior extensively. This 226 // is simply a short, additional test. 227 TEST(UTFOffsetStringConversionsTest, MergeSequentialAdjustments) { 228 // Pretend the input string is "abcdefghijklmnopqrstuvwxyz". 229 230 // Set up |first_adjustments| to 231 // - remove the leading "a" 232 // - combine the "bc" into one character (call it ".") 233 // - remove the "f" 234 // - remove the "tuv" 235 // The resulting string should be ".deghijklmnopqrswxyz". 236 OffsetAdjuster::Adjustments first_adjustments; 237 first_adjustments.push_back(OffsetAdjuster::Adjustment(0, 1, 0)); 238 first_adjustments.push_back(OffsetAdjuster::Adjustment(1, 2, 1)); 239 first_adjustments.push_back(OffsetAdjuster::Adjustment(5, 1, 0)); 240 first_adjustments.push_back(OffsetAdjuster::Adjustment(19, 3, 0)); 241 242 // Set up |adjustments_on_adjusted_string| to 243 // - combine the "." character that replaced "bc" with "d" into one character 244 // (call it "?") 245 // - remove the "egh" 246 // - expand the "i" into two characters (call them "12") 247 // - combine the "jkl" into one character (call it "@") 248 // - expand the "z" into two characters (call it "34") 249 // The resulting string should be "?12@mnopqrswxy34". 250 OffsetAdjuster::Adjustments adjustments_on_adjusted_string; 251 adjustments_on_adjusted_string.push_back(OffsetAdjuster::Adjustment( 252 0, 2, 1)); 253 adjustments_on_adjusted_string.push_back(OffsetAdjuster::Adjustment( 254 2, 3, 0)); 255 adjustments_on_adjusted_string.push_back(OffsetAdjuster::Adjustment( 256 5, 1, 2)); 257 adjustments_on_adjusted_string.push_back(OffsetAdjuster::Adjustment( 258 6, 3, 1)); 259 adjustments_on_adjusted_string.push_back(OffsetAdjuster::Adjustment( 260 19, 1, 2)); 261 262 // Now merge the adjustments and check the results. 263 OffsetAdjuster::MergeSequentialAdjustments(first_adjustments, 264 &adjustments_on_adjusted_string); 265 // The merged adjustments should look like 266 // - combine abcd into "?" 267 // - note: it's also reasonable for the Merge function to instead produce 268 // two adjustments instead of this, one to remove a and another to 269 // combine bcd into "?". This test verifies the current behavior. 270 // - remove efgh 271 // - expand i into "12" 272 // - combine jkl into "@" 273 // - remove tuv 274 // - expand z into "34" 275 ASSERT_EQ(6u, adjustments_on_adjusted_string.size()); 276 EXPECT_EQ(0u, adjustments_on_adjusted_string[0].original_offset); 277 EXPECT_EQ(4u, adjustments_on_adjusted_string[0].original_length); 278 EXPECT_EQ(1u, adjustments_on_adjusted_string[0].output_length); 279 EXPECT_EQ(4u, adjustments_on_adjusted_string[1].original_offset); 280 EXPECT_EQ(4u, adjustments_on_adjusted_string[1].original_length); 281 EXPECT_EQ(0u, adjustments_on_adjusted_string[1].output_length); 282 EXPECT_EQ(8u, adjustments_on_adjusted_string[2].original_offset); 283 EXPECT_EQ(1u, adjustments_on_adjusted_string[2].original_length); 284 EXPECT_EQ(2u, adjustments_on_adjusted_string[2].output_length); 285 EXPECT_EQ(9u, adjustments_on_adjusted_string[3].original_offset); 286 EXPECT_EQ(3u, adjustments_on_adjusted_string[3].original_length); 287 EXPECT_EQ(1u, adjustments_on_adjusted_string[3].output_length); 288 EXPECT_EQ(19u, adjustments_on_adjusted_string[4].original_offset); 289 EXPECT_EQ(3u, adjustments_on_adjusted_string[4].original_length); 290 EXPECT_EQ(0u, adjustments_on_adjusted_string[4].output_length); 291 EXPECT_EQ(25u, adjustments_on_adjusted_string[5].original_offset); 292 EXPECT_EQ(1u, adjustments_on_adjusted_string[5].original_length); 293 EXPECT_EQ(2u, adjustments_on_adjusted_string[5].output_length); 294 } 295 296 } // namaspace base 297