1 /* 2 * Copyright (C) 2014 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 #include "utils/char_utils.h" 18 19 #include <gtest/gtest.h> 20 21 #include "defines.h" 22 23 namespace latinime { 24 namespace { 25 26 TEST(CharUtilsTest, TestIsAsciiUpper) { 27 EXPECT_TRUE(CharUtils::isAsciiUpper('A')); 28 EXPECT_TRUE(CharUtils::isAsciiUpper('Z')); 29 EXPECT_FALSE(CharUtils::isAsciiUpper('a')); 30 EXPECT_FALSE(CharUtils::isAsciiUpper('z')); 31 EXPECT_FALSE(CharUtils::isAsciiUpper('@')); 32 EXPECT_FALSE(CharUtils::isAsciiUpper(' ')); 33 EXPECT_FALSE(CharUtils::isAsciiUpper(0x00C0 /* LATIN CAPITAL LETTER A WITH GRAVE */)); 34 EXPECT_FALSE(CharUtils::isAsciiUpper(0x00E0 /* LATIN SMALL LETTER A WITH GRAVE */)); 35 EXPECT_FALSE(CharUtils::isAsciiUpper(0x03C2 /* GREEK SMALL LETTER FINAL SIGMA */)); 36 EXPECT_FALSE(CharUtils::isAsciiUpper(0x0410 /* CYRILLIC CAPITAL LETTER A */)); 37 EXPECT_FALSE(CharUtils::isAsciiUpper(0x0430 /* CYRILLIC SMALL LETTER A */)); 38 EXPECT_FALSE(CharUtils::isAsciiUpper(0x3042 /* HIRAGANA LETTER A */)); 39 EXPECT_FALSE(CharUtils::isAsciiUpper(0x1F36A /* COOKIE */)); 40 } 41 42 TEST(CharUtilsTest, TestToLowerCase) { 43 EXPECT_EQ('a', CharUtils::toLowerCase('A')); 44 EXPECT_EQ('z', CharUtils::toLowerCase('Z')); 45 EXPECT_EQ('a', CharUtils::toLowerCase('a')); 46 EXPECT_EQ('z', CharUtils::toLowerCase('z')); 47 EXPECT_EQ('@', CharUtils::toLowerCase('@')); 48 EXPECT_EQ(' ', CharUtils::toLowerCase(' ')); 49 EXPECT_EQ(0x00E0 /* LATIN SMALL LETTER A WITH GRAVE */, 50 CharUtils::toLowerCase(0x00C0 /* LATIN CAPITAL LETTER A WITH GRAVE */)); 51 EXPECT_EQ(0x00E0 /* LATIN SMALL LETTER A WITH GRAVE */, 52 CharUtils::toLowerCase(0x00E0 /* LATIN SMALL LETTER A WITH GRAVE */)); 53 EXPECT_EQ(0x03C2 /* GREEK SMALL LETTER FINAL SIGMA */, 54 CharUtils::toLowerCase(0x03C2 /* GREEK SMALL LETTER FINAL SIGMA */)); 55 EXPECT_EQ(0x0430 /* CYRILLIC SMALL LETTER A */, 56 CharUtils::toLowerCase(0x0410 /* CYRILLIC CAPITAL LETTER A */)); 57 EXPECT_EQ(0x0430 /* CYRILLIC SMALL LETTER A */, 58 CharUtils::toLowerCase(0x0430 /* CYRILLIC SMALL LETTER A */)); 59 EXPECT_EQ(0x3042 /* HIRAGANA LETTER A */, 60 CharUtils::toLowerCase(0x3042 /* HIRAGANA LETTER A */)); 61 EXPECT_EQ(0x1F36A /* COOKIE */, CharUtils::toLowerCase(0x1F36A /* COOKIE */)); 62 } 63 64 TEST(CharUtilsTest, TestToBaseLowerCase) { 65 EXPECT_EQ('a', CharUtils::toBaseLowerCase('A')); 66 EXPECT_EQ('z', CharUtils::toBaseLowerCase('Z')); 67 EXPECT_EQ('a', CharUtils::toBaseLowerCase('a')); 68 EXPECT_EQ('z', CharUtils::toBaseLowerCase('z')); 69 EXPECT_EQ('@', CharUtils::toBaseLowerCase('@')); 70 EXPECT_EQ(' ', CharUtils::toBaseLowerCase(' ')); 71 EXPECT_EQ('a', CharUtils::toBaseLowerCase(0x00C0 /* LATIN CAPITAL LETTER A WITH GRAVE */)); 72 EXPECT_EQ('a', CharUtils::toBaseLowerCase(0x00E0 /* LATIN SMALL LETTER A WITH GRAVE */)); 73 EXPECT_EQ(0x03C2 /* GREEK SMALL LETTER FINAL SIGMA */, 74 CharUtils::toBaseLowerCase(0x03C2 /* GREEK SMALL LETTER FINAL SIGMA */)); 75 EXPECT_EQ(0x0430 /* CYRILLIC SMALL LETTER A */, 76 CharUtils::toBaseLowerCase(0x0410 /* CYRILLIC CAPITAL LETTER A */)); 77 EXPECT_EQ(0x0430 /* CYRILLIC SMALL LETTER A */, 78 CharUtils::toBaseLowerCase(0x0430 /* CYRILLIC SMALL LETTER A */)); 79 EXPECT_EQ(0x3042 /* HIRAGANA LETTER A */, 80 CharUtils::toBaseLowerCase(0x3042 /* HIRAGANA LETTER A */)); 81 EXPECT_EQ(0x1F36A /* COOKIE */, CharUtils::toBaseLowerCase(0x1F36A /* COOKIE */)); 82 } 83 84 TEST(CharUtilsTest, TestToBaseCodePoint) { 85 EXPECT_EQ('A', CharUtils::toBaseCodePoint('A')); 86 EXPECT_EQ('Z', CharUtils::toBaseCodePoint('Z')); 87 EXPECT_EQ('a', CharUtils::toBaseCodePoint('a')); 88 EXPECT_EQ('z', CharUtils::toBaseCodePoint('z')); 89 EXPECT_EQ('@', CharUtils::toBaseCodePoint('@')); 90 EXPECT_EQ(' ', CharUtils::toBaseCodePoint(' ')); 91 EXPECT_EQ('A', CharUtils::toBaseCodePoint(0x00C0 /* LATIN CAPITAL LETTER A WITH GRAVE */)); 92 EXPECT_EQ('a', CharUtils::toBaseCodePoint(0x00E0 /* LATIN SMALL LETTER A WITH GRAVE */)); 93 EXPECT_EQ(0x03C2 /* GREEK SMALL LETTER FINAL SIGMA */, 94 CharUtils::toBaseLowerCase(0x03C2 /* GREEK SMALL LETTER FINAL SIGMA */)); 95 EXPECT_EQ(0x0410 /* CYRILLIC CAPITAL LETTER A */, 96 CharUtils::toBaseCodePoint(0x0410 /* CYRILLIC CAPITAL LETTER A */)); 97 EXPECT_EQ(0x0430 /* CYRILLIC SMALL LETTER A */, 98 CharUtils::toBaseCodePoint(0x0430 /* CYRILLIC SMALL LETTER A */)); 99 EXPECT_EQ(0x3042 /* HIRAGANA LETTER A */, 100 CharUtils::toBaseCodePoint(0x3042 /* HIRAGANA LETTER A */)); 101 EXPECT_EQ(0x1F36A /* COOKIE */, CharUtils::toBaseCodePoint(0x1F36A /* COOKIE */)); 102 } 103 104 TEST(CharUtilsTest, TestIsIntentionalOmissionCodePoint) { 105 EXPECT_TRUE(CharUtils::isIntentionalOmissionCodePoint('\'')); 106 EXPECT_TRUE(CharUtils::isIntentionalOmissionCodePoint('-')); 107 EXPECT_FALSE(CharUtils::isIntentionalOmissionCodePoint('a')); 108 EXPECT_FALSE(CharUtils::isIntentionalOmissionCodePoint('?')); 109 EXPECT_FALSE(CharUtils::isIntentionalOmissionCodePoint('/')); 110 } 111 112 TEST(CharUtilsTest, TestIsInUnicodeSpace) { 113 EXPECT_FALSE(CharUtils::isInUnicodeSpace(NOT_A_CODE_POINT)); 114 EXPECT_FALSE(CharUtils::isInUnicodeSpace(CODE_POINT_BEGINNING_OF_SENTENCE)); 115 EXPECT_TRUE(CharUtils::isInUnicodeSpace('a')); 116 EXPECT_TRUE(CharUtils::isInUnicodeSpace(0x0410 /* CYRILLIC CAPITAL LETTER A */)); 117 EXPECT_TRUE(CharUtils::isInUnicodeSpace(0x3042 /* HIRAGANA LETTER A */)); 118 EXPECT_TRUE(CharUtils::isInUnicodeSpace(0x1F36A /* COOKIE */)); 119 } 120 121 } // namespace 122 } // namespace latinime 123
