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 package java.lang; 18 19 import android.icu.text.Transliterator; 20 import java.util.Locale; 21 import libcore.icu.ICU; 22 23 /** 24 * Performs case operations as described by http://unicode.org/reports/tr21/tr21-5.html. 25 */ 26 class CaseMapper { 27 private static final char[] upperValues = "SS\u0000\u02bcN\u0000J\u030c\u0000\u0399\u0308\u0301\u03a5\u0308\u0301\u0535\u0552\u0000H\u0331\u0000T\u0308\u0000W\u030a\u0000Y\u030a\u0000A\u02be\u0000\u03a5\u0313\u0000\u03a5\u0313\u0300\u03a5\u0313\u0301\u03a5\u0313\u0342\u1f08\u0399\u0000\u1f09\u0399\u0000\u1f0a\u0399\u0000\u1f0b\u0399\u0000\u1f0c\u0399\u0000\u1f0d\u0399\u0000\u1f0e\u0399\u0000\u1f0f\u0399\u0000\u1f08\u0399\u0000\u1f09\u0399\u0000\u1f0a\u0399\u0000\u1f0b\u0399\u0000\u1f0c\u0399\u0000\u1f0d\u0399\u0000\u1f0e\u0399\u0000\u1f0f\u0399\u0000\u1f28\u0399\u0000\u1f29\u0399\u0000\u1f2a\u0399\u0000\u1f2b\u0399\u0000\u1f2c\u0399\u0000\u1f2d\u0399\u0000\u1f2e\u0399\u0000\u1f2f\u0399\u0000\u1f28\u0399\u0000\u1f29\u0399\u0000\u1f2a\u0399\u0000\u1f2b\u0399\u0000\u1f2c\u0399\u0000\u1f2d\u0399\u0000\u1f2e\u0399\u0000\u1f2f\u0399\u0000\u1f68\u0399\u0000\u1f69\u0399\u0000\u1f6a\u0399\u0000\u1f6b\u0399\u0000\u1f6c\u0399\u0000\u1f6d\u0399\u0000\u1f6e\u0399\u0000\u1f6f\u0399\u0000\u1f68\u0399\u0000\u1f69\u0399\u0000\u1f6a\u0399\u0000\u1f6b\u0399\u0000\u1f6c\u0399\u0000\u1f6d\u0399\u0000\u1f6e\u0399\u0000\u1f6f\u0399\u0000\u1fba\u0399\u0000\u0391\u0399\u0000\u0386\u0399\u0000\u0391\u0342\u0000\u0391\u0342\u0399\u0391\u0399\u0000\u1fca\u0399\u0000\u0397\u0399\u0000\u0389\u0399\u0000\u0397\u0342\u0000\u0397\u0342\u0399\u0397\u0399\u0000\u0399\u0308\u0300\u0399\u0308\u0301\u0399\u0342\u0000\u0399\u0308\u0342\u03a5\u0308\u0300\u03a5\u0308\u0301\u03a1\u0313\u0000\u03a5\u0342\u0000\u03a5\u0308\u0342\u1ffa\u0399\u0000\u03a9\u0399\u0000\u038f\u0399\u0000\u03a9\u0342\u0000\u03a9\u0342\u0399\u03a9\u0399\u0000FF\u0000FI\u0000FL\u0000FFIFFLST\u0000ST\u0000\u0544\u0546\u0000\u0544\u0535\u0000\u0544\u053b\u0000\u054e\u0546\u0000\u0544\u053d\u0000".toCharArray(); 28 private static final char[] upperValues2 = "\u000b\u0000\f\u0000\r\u0000\u000e\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u000f\u0010\u0011\u0012\u0013\u0014\u0015\u0016\u0017\u0018\u0019\u001a\u001b\u001c\u001d\u001e\u001f !\"#$%&'()*+,-./0123456789:;<=>\u0000\u0000?@A\u0000BC\u0000\u0000\u0000\u0000D\u0000\u0000\u0000\u0000\u0000EFG\u0000HI\u0000\u0000\u0000\u0000J\u0000\u0000\u0000\u0000\u0000KL\u0000\u0000MN\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000OPQ\u0000RS\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000\u0000TUV\u0000WX\u0000\u0000\u0000\u0000Y".toCharArray(); 29 30 private static final char LATIN_CAPITAL_I_WITH_DOT = '\u0130'; 31 private static final char GREEK_CAPITAL_SIGMA = '\u03a3'; 32 private static final char GREEK_SMALL_FINAL_SIGMA = '\u03c2'; 33 34 /** 35 * Our current GC makes short-lived objects more expensive than we'd like. When that's fixed, 36 * this class should be changed so that you instantiate it with the String and its value, 37 * and count fields. 38 */ CaseMapper()39 private CaseMapper() { 40 } 41 42 /** 43 * Implements String.toLowerCase. The original String instance is returned if nothing changes. 44 */ toLowerCase(Locale locale, String s)45 public static String toLowerCase(Locale locale, String s) { 46 // Punt hard cases to ICU4C. 47 // Note that Greek isn't a particularly hard case for toLowerCase, only toUpperCase. 48 String languageCode = locale.getLanguage(); 49 if (languageCode.equals("tr") || languageCode.equals("az") || languageCode.equals("lt")) { 50 return ICU.toLowerCase(s, locale); 51 } 52 53 String newString = null; 54 for (int i = 0, end = s.length(); i < end; ++i) { 55 char ch = s.charAt(i); 56 char newCh; 57 if (ch == LATIN_CAPITAL_I_WITH_DOT || Character.isHighSurrogate(ch)) { 58 // Punt these hard cases. 59 return ICU.toLowerCase(s, locale); 60 } else if (ch == GREEK_CAPITAL_SIGMA && isFinalSigma(s, i)) { 61 newCh = GREEK_SMALL_FINAL_SIGMA; 62 } else { 63 newCh = Character.toLowerCase(ch); 64 } 65 if (ch != newCh) { 66 if (newString == null) { 67 newString = StringFactory.newStringFromString(s); 68 } 69 newString.setCharAt(i, newCh); 70 } 71 } 72 return newString != null ? newString : s; 73 } 74 75 /** 76 * True if 'index' is preceded by a sequence consisting of a cased letter and a case-ignorable 77 * sequence, and 'index' is not followed by a sequence consisting of an ignorable sequence and 78 * then a cased letter. 79 */ isFinalSigma(String s, int index)80 private static boolean isFinalSigma(String s, int index) { 81 // TODO: we don't skip case-ignorable sequences like we should. 82 // TODO: we should add a more direct way to test for a cased letter. 83 if (index <= 0) { 84 return false; 85 } 86 char previous = s.charAt(index - 1); 87 if (!(Character.isLowerCase(previous) || Character.isUpperCase(previous) || Character.isTitleCase(previous))) { 88 return false; 89 } 90 if (index + 1 >= s.length()) { 91 return true; 92 } 93 char next = s.charAt(index + 1); 94 if (Character.isLowerCase(next) || Character.isUpperCase(next) || Character.isTitleCase(next)) { 95 return false; 96 } 97 return true; 98 } 99 100 /** 101 * Return the index of the specified character into the upperValues table. 102 * The upperValues table contains three entries at each position. These 103 * three characters are the upper case conversion. If only two characters 104 * are used, the third character in the table is \u0000. 105 * @return the index into the upperValues table, or -1 106 */ upperIndex(int ch)107 private static int upperIndex(int ch) { 108 int index = -1; 109 if (ch >= 0xdf) { 110 if (ch <= 0x587) { 111 switch (ch) { 112 case 0xdf: return 0; 113 case 0x149: return 1; 114 case 0x1f0: return 2; 115 case 0x390: return 3; 116 case 0x3b0: return 4; 117 case 0x587: return 5; 118 } 119 } else if (ch >= 0x1e96) { 120 if (ch <= 0x1e9a) { 121 index = 6 + ch - 0x1e96; 122 } else if (ch >= 0x1f50 && ch <= 0x1ffc) { 123 index = upperValues2[ch - 0x1f50]; 124 if (index == 0) { 125 index = -1; 126 } 127 } else if (ch >= 0xfb00) { 128 if (ch <= 0xfb06) { 129 index = 90 + ch - 0xfb00; 130 } else if (ch >= 0xfb13 && ch <= 0xfb17) { 131 index = 97 + ch - 0xfb13; 132 } 133 } 134 } 135 } 136 return index; 137 } 138 139 private static final ThreadLocal<Transliterator> EL_UPPER = new ThreadLocal<Transliterator>() { 140 @Override protected Transliterator initialValue() { 141 return Transliterator.getInstance("el-Upper"); 142 } 143 }; 144 toUpperCase(Locale locale, String s, int count)145 public static String toUpperCase(Locale locale, String s, int count) { 146 String languageCode = locale.getLanguage(); 147 if (languageCode.equals("tr") || languageCode.equals("az") || languageCode.equals("lt")) { 148 return ICU.toUpperCase(s, locale); 149 } 150 if (languageCode.equals("el")) { 151 return EL_UPPER.get().transliterate(s); 152 } 153 154 char[] output = null; 155 String newString = null; 156 int i = 0; 157 for (int o = 0, end = count; o < end; o++) { 158 char ch = s.charAt(o); 159 if (Character.isHighSurrogate(ch)) { 160 return ICU.toUpperCase(s, locale); 161 } 162 int index = upperIndex(ch); 163 if (index == -1) { 164 if (output != null && i >= output.length) { 165 char[] newoutput = new char[output.length + (count / 6) + 2]; 166 System.arraycopy(output, 0, newoutput, 0, output.length); 167 output = newoutput; 168 } 169 char upch = Character.toUpperCase(ch); 170 if (output != null) { 171 output[i++] = upch; 172 } else if (ch != upch) { 173 if (newString == null) { 174 newString = StringFactory.newStringFromString(s); 175 } 176 newString.setCharAt(o, upch); 177 } 178 } else { 179 int target = index * 3; 180 char val3 = upperValues[target + 2]; 181 if (output == null) { 182 output = new char[count + (count / 6) + 2]; 183 i = o; 184 if (newString != null) { 185 System.arraycopy(newString.toCharArray(), 0, output, 0, i); 186 } else { 187 System.arraycopy(s.toCharArray(), 0, output, 0, i); 188 } 189 } else if (i + (val3 == 0 ? 1 : 2) >= output.length) { 190 char[] newoutput = new char[output.length + (count / 6) + 3]; 191 System.arraycopy(output, 0, newoutput, 0, output.length); 192 output = newoutput; 193 } 194 195 char val = upperValues[target]; 196 output[i++] = val; 197 val = upperValues[target + 1]; 198 output[i++] = val; 199 if (val3 != 0) { 200 output[i++] = val3; 201 } 202 } 203 } 204 if (output == null) { 205 if (newString != null) { 206 return newString; 207 } else { 208 return s; 209 } 210 } 211 return output.length == i || output.length - i < 8 ? new String(0, i, output) : new String(output, 0, i); 212 } 213 } 214