1// © 2016 and later: Unicode, Inc. and others. 2// License & terms of use: http://www.unicode.org/copyright.html#License 3/** 4******************************************************************************* 5* Copyright (C) 1996-2010, International Business Machines Corporation and * 6* others. All Rights Reserved. * 7******************************************************************************* 8*/ 9 10package com.ibm.icu.dev.demo.translit; 11import java.util.HashMap; 12import java.util.HashSet; 13import java.util.Iterator; 14import java.util.Map; 15import java.util.Set; 16import java.util.TreeSet; 17 18import com.ibm.icu.lang.UCharacter; 19import com.ibm.icu.text.Transliterator; 20import com.ibm.icu.text.UTF16; 21import com.ibm.icu.text.UnicodeSet; 22 23/** 24 * Incrementally returns the set of all strings that case-fold to the same value. 25 */ 26public class CaseIterator { 27 28 // testing stuff 29 static Transliterator toName = Transliterator.getInstance("[:^ascii:] Any-Name"); 30 static Transliterator toHex = Transliterator.getInstance("[:^ascii:] Any-Hex"); 31 static Transliterator toHex2 = Transliterator.getInstance("[[^\u0021-\u007F]-[,]] Any-Hex"); 32 33 // global tables (could be precompiled) 34 private static Map fromCaseFold = new HashMap(); 35 private static Map toCaseFold = new HashMap(); 36 private static int maxLength = 0; 37 38 // This exception list is generated on the console by turning on the GENERATED flag, 39 // which MUST be false for normal operation. 40 // Once the list is generated, it is pasted in here. 41 // A bit of a cludge, but this bootstrapping is the easiest way 42 // to get around certain complications in the data. 43 44 private static final boolean GENERATE = false; 45 46 private static final boolean DUMP = false; 47 48 private static String[][] exceptionList = { 49 // a\N{MODIFIER LETTER RIGHT HALF RING} 50 {"a\u02BE","A\u02BE","a\u02BE",}, 51 // ff 52 {"ff","FF","Ff","fF","ff",}, 53 // ffi 54 {"ffi","FFI","FFi","FfI","Ffi","F\uFB01","fFI","fFi","ffI","ffi","f\uFB01","\uFB00I","\uFB00i",}, 55 // ffl 56 {"ffl","FFL","FFl","FfL","Ffl","F\uFB02","fFL","fFl","ffL","ffl","f\uFB02","\uFB00L","\uFB00l",}, 57 // fi 58 {"fi","FI","Fi","fI","fi",}, 59 // fl 60 {"fl","FL","Fl","fL","fl",}, 61 // h\N{COMBINING MACRON BELOW} 62 {"h\u0331","H\u0331","h\u0331",}, 63 // i\N{COMBINING DOT ABOVE} 64 {"i\u0307","I\u0307","i\u0307",}, 65 // j\N{COMBINING CARON} 66 {"j\u030C","J\u030C","j\u030C",}, 67 // ss 68 {"ss","SS","Ss","S\u017F","sS","ss","s\u017F","\u017FS","\u017Fs","\u017F\u017F",}, 69 // st 70 {"st","ST","St","sT","st","\u017FT","\u017Ft",}, 71 // t\N{COMBINING DIAERESIS} 72 {"t\u0308","T\u0308","t\u0308",}, 73 // w\N{COMBINING RING ABOVE} 74 {"w\u030A","W\u030A","w\u030A",}, 75 // y\N{COMBINING RING ABOVE} 76 {"y\u030A","Y\u030A","y\u030A",}, 77 // \N{MODIFIER LETTER APOSTROPHE}n 78 {"\u02BCn","\u02BCN","\u02BCn",}, 79 // \N{GREEK SMALL LETTER ALPHA WITH TONOS}\N{GREEK SMALL LETTER IOTA} 80 {"\u03AC\u03B9","\u0386\u0345","\u0386\u0399","\u0386\u03B9","\u0386\u1FBE","\u03AC\u0345","\u03AC\u0399","\u03AC\u03B9","\u03AC\u1FBE",}, 81 // \N{GREEK SMALL LETTER ETA WITH TONOS}\N{GREEK SMALL LETTER IOTA} 82 {"\u03AE\u03B9","\u0389\u0345","\u0389\u0399","\u0389\u03B9","\u0389\u1FBE","\u03AE\u0345","\u03AE\u0399","\u03AE\u03B9","\u03AE\u1FBE",}, 83 // \N{GREEK SMALL LETTER ALPHA}\N{COMBINING GREEK PERISPOMENI} 84 {"\u03B1\u0342","\u0391\u0342","\u03B1\u0342",}, 85 // \N{GREEK SMALL LETTER ALPHA}\N{COMBINING GREEK PERISPOMENI}\N{GREEK SMALL LETTER IOTA} 86 {"\u03B1\u0342\u03B9","\u0391\u0342\u0345","\u0391\u0342\u0399","\u0391\u0342\u03B9","\u0391\u0342\u1FBE", 87 "\u03B1\u0342\u0345","\u03B1\u0342\u0399","\u03B1\u0342\u03B9","\u03B1\u0342\u1FBE","\u1FB6\u0345", 88 "\u1FB6\u0399","\u1FB6\u03B9","\u1FB6\u1FBE",}, 89 // \N{GREEK SMALL LETTER ALPHA}\N{GREEK SMALL LETTER IOTA} 90 {"\u03B1\u03B9","\u0391\u0345","\u0391\u0399","\u0391\u03B9","\u0391\u1FBE","\u03B1\u0345","\u03B1\u0399","\u03B1\u03B9","\u03B1\u1FBE",}, 91 // \N{GREEK SMALL LETTER ETA}\N{COMBINING GREEK PERISPOMENI} 92 {"\u03B7\u0342","\u0397\u0342","\u03B7\u0342",}, 93 // \N{GREEK SMALL LETTER ETA}\N{COMBINING GREEK PERISPOMENI}\N{GREEK SMALL LETTER IOTA} 94 {"\u03B7\u0342\u03B9","\u0397\u0342\u0345","\u0397\u0342\u0399","\u0397\u0342\u03B9","\u0397\u0342\u1FBE", 95 "\u03B7\u0342\u0345","\u03B7\u0342\u0399","\u03B7\u0342\u03B9","\u03B7\u0342\u1FBE","\u1FC6\u0345","\u1FC6\u0399", 96 "\u1FC6\u03B9","\u1FC6\u1FBE",}, 97 // \N{GREEK SMALL LETTER ETA}\N{GREEK SMALL LETTER IOTA} 98 {"\u03B7\u03B9","\u0397\u0345","\u0397\u0399","\u0397\u03B9","\u0397\u1FBE","\u03B7\u0345","\u03B7\u0399","\u03B7\u03B9","\u03B7\u1FBE",}, 99 // \N{GREEK SMALL LETTER IOTA}\N{COMBINING DIAERESIS}\N{COMBINING GRAVE ACCENT} 100 {"\u03B9\u0308\u0300","\u0345\u0308\u0300","\u0399\u0308\u0300","\u03B9\u0308\u0300","\u1FBE\u0308\u0300",}, 101 // \N{GREEK SMALL LETTER IOTA}\N{COMBINING DIAERESIS}\N{COMBINING ACUTE ACCENT} 102 {"\u03B9\u0308\u0301","\u0345\u0308\u0301","\u0399\u0308\u0301","\u03B9\u0308\u0301","\u1FBE\u0308\u0301",}, 103 // \N{GREEK SMALL LETTER IOTA}\N{COMBINING DIAERESIS}\N{COMBINING GREEK PERISPOMENI} 104 {"\u03B9\u0308\u0342","\u0345\u0308\u0342","\u0399\u0308\u0342","\u03B9\u0308\u0342","\u1FBE\u0308\u0342",}, 105 // \N{GREEK SMALL LETTER IOTA}\N{COMBINING GREEK PERISPOMENI} 106 {"\u03B9\u0342","\u0345\u0342","\u0399\u0342","\u03B9\u0342","\u1FBE\u0342",}, 107 // \N{GREEK SMALL LETTER RHO}\N{COMBINING COMMA ABOVE} 108 {"\u03C1\u0313","\u03A1\u0313","\u03C1\u0313","\u03F1\u0313",}, 109 // \N{GREEK SMALL LETTER UPSILON}\N{COMBINING DIAERESIS}\N{COMBINING GRAVE ACCENT} 110 {"\u03C5\u0308\u0300","\u03A5\u0308\u0300","\u03C5\u0308\u0300",}, 111 // \N{GREEK SMALL LETTER UPSILON}\N{COMBINING DIAERESIS}\N{COMBINING ACUTE ACCENT} 112 {"\u03C5\u0308\u0301","\u03A5\u0308\u0301","\u03C5\u0308\u0301",}, 113 // \N{GREEK SMALL LETTER UPSILON}\N{COMBINING DIAERESIS}\N{COMBINING GREEK PERISPOMENI} 114 {"\u03C5\u0308\u0342","\u03A5\u0308\u0342","\u03C5\u0308\u0342",}, 115 // \N{GREEK SMALL LETTER UPSILON}\N{COMBINING COMMA ABOVE} 116 {"\u03C5\u0313","\u03A5\u0313","\u03C5\u0313",}, 117 // \N{GREEK SMALL LETTER UPSILON}\N{COMBINING COMMA ABOVE}\N{COMBINING GRAVE ACCENT} 118 {"\u03C5\u0313\u0300","\u03A5\u0313\u0300","\u03C5\u0313\u0300","\u1F50\u0300",}, 119 // \N{GREEK SMALL LETTER UPSILON}\N{COMBINING COMMA ABOVE}\N{COMBINING ACUTE ACCENT} 120 {"\u03C5\u0313\u0301","\u03A5\u0313\u0301","\u03C5\u0313\u0301","\u1F50\u0301",}, 121 // \N{GREEK SMALL LETTER UPSILON}\N{COMBINING COMMA ABOVE}\N{COMBINING GREEK PERISPOMENI} 122 {"\u03C5\u0313\u0342","\u03A5\u0313\u0342","\u03C5\u0313\u0342","\u1F50\u0342",}, 123 // \N{GREEK SMALL LETTER UPSILON}\N{COMBINING GREEK PERISPOMENI} 124 {"\u03C5\u0342","\u03A5\u0342","\u03C5\u0342",}, 125 // \N{GREEK SMALL LETTER OMEGA}\N{COMBINING GREEK PERISPOMENI} 126 {"\u03C9\u0342","\u03A9\u0342","\u03C9\u0342","\u2126\u0342",}, 127 // \N{GREEK SMALL LETTER OMEGA}\N{COMBINING GREEK PERISPOMENI}\N{GREEK SMALL LETTER IOTA} 128 {"\u03C9\u0342\u03B9","\u03A9\u0342\u0345","\u03A9\u0342\u0399","\u03A9\u0342\u03B9","\u03A9\u0342\u1FBE","\u03C9\u0342\u0345","\u03C9\u0342\u0399","\u03C9\u0342\u03B9","\u03C9\u0342\u1FBE","\u1FF6\u0345", 129 "\u1FF6\u0399","\u1FF6\u03B9","\u1FF6\u1FBE","\u2126\u0342\u0345","\u2126\u0342\u0399","\u2126\u0342\u03B9","\u2126\u0342\u1FBE",}, 130 // \N{GREEK SMALL LETTER OMEGA}\N{GREEK SMALL LETTER IOTA} 131 {"\u03C9\u03B9","\u03A9\u0345","\u03A9\u0399","\u03A9\u03B9","\u03A9\u1FBE","\u03C9\u0345","\u03C9\u0399","\u03C9\u03B9","\u03C9\u1FBE","\u2126\u0345","\u2126\u0399","\u2126\u03B9","\u2126\u1FBE",}, 132 // \N{GREEK SMALL LETTER OMEGA WITH TONOS}\N{GREEK SMALL LETTER IOTA} 133 {"\u03CE\u03B9","\u038F\u0345","\u038F\u0399","\u038F\u03B9","\u038F\u1FBE","\u03CE\u0345","\u03CE\u0399","\u03CE\u03B9","\u03CE\u1FBE",}, 134 // \N{ARMENIAN SMALL LETTER ECH}\N{ARMENIAN SMALL LETTER YIWN} 135 {"\u0565\u0582","\u0535\u0552","\u0535\u0582","\u0565\u0552","\u0565\u0582",}, 136 // \N{ARMENIAN SMALL LETTER MEN}\N{ARMENIAN SMALL LETTER ECH} 137 {"\u0574\u0565","\u0544\u0535","\u0544\u0565","\u0574\u0535","\u0574\u0565",}, 138 // \N{ARMENIAN SMALL LETTER MEN}\N{ARMENIAN SMALL LETTER INI} 139 {"\u0574\u056B","\u0544\u053B","\u0544\u056B","\u0574\u053B","\u0574\u056B",}, 140 // \N{ARMENIAN SMALL LETTER MEN}\N{ARMENIAN SMALL LETTER XEH} 141 {"\u0574\u056D","\u0544\u053D","\u0544\u056D","\u0574\u053D","\u0574\u056D",}, 142 // \N{ARMENIAN SMALL LETTER MEN}\N{ARMENIAN SMALL LETTER NOW} 143 {"\u0574\u0576","\u0544\u0546","\u0544\u0576","\u0574\u0546","\u0574\u0576",}, 144 // \N{ARMENIAN SMALL LETTER VEW}\N{ARMENIAN SMALL LETTER NOW} 145 {"\u057E\u0576","\u054E\u0546","\u054E\u0576","\u057E\u0546","\u057E\u0576",}, 146 // \N{GREEK SMALL LETTER ALPHA WITH PSILI}\N{GREEK SMALL LETTER IOTA} 147 {"\u1F00\u03B9","\u1F00\u0345","\u1F00\u0399","\u1F00\u03B9","\u1F00\u1FBE","\u1F08\u0345","\u1F08\u0399","\u1F08\u03B9","\u1F08\u1FBE",}, 148 // \N{GREEK SMALL LETTER ALPHA WITH DASIA}\N{GREEK SMALL LETTER IOTA} 149 {"\u1F01\u03B9","\u1F01\u0345","\u1F01\u0399","\u1F01\u03B9","\u1F01\u1FBE","\u1F09\u0345","\u1F09\u0399","\u1F09\u03B9","\u1F09\u1FBE",}, 150 // \N{GREEK SMALL LETTER ALPHA WITH PSILI AND VARIA}\N{GREEK SMALL LETTER IOTA} 151 {"\u1F02\u03B9","\u1F02\u0345","\u1F02\u0399","\u1F02\u03B9","\u1F02\u1FBE","\u1F0A\u0345","\u1F0A\u0399","\u1F0A\u03B9","\u1F0A\u1FBE",}, 152 // \N{GREEK SMALL LETTER ALPHA WITH DASIA AND VARIA}\N{GREEK SMALL LETTER IOTA} 153 {"\u1F03\u03B9","\u1F03\u0345","\u1F03\u0399","\u1F03\u03B9","\u1F03\u1FBE","\u1F0B\u0345","\u1F0B\u0399","\u1F0B\u03B9","\u1F0B\u1FBE",}, 154 // \N{GREEK SMALL LETTER ALPHA WITH PSILI AND OXIA}\N{GREEK SMALL LETTER IOTA} 155 {"\u1F04\u03B9","\u1F04\u0345","\u1F04\u0399","\u1F04\u03B9","\u1F04\u1FBE","\u1F0C\u0345","\u1F0C\u0399","\u1F0C\u03B9","\u1F0C\u1FBE",}, 156 // \N{GREEK SMALL LETTER ALPHA WITH DASIA AND OXIA}\N{GREEK SMALL LETTER IOTA} 157 {"\u1F05\u03B9","\u1F05\u0345","\u1F05\u0399","\u1F05\u03B9","\u1F05\u1FBE","\u1F0D\u0345","\u1F0D\u0399","\u1F0D\u03B9","\u1F0D\u1FBE",}, 158 // \N{GREEK SMALL LETTER ALPHA WITH PSILI AND PERISPOMENI}\N{GREEK SMALL LETTER IOTA} 159 {"\u1F06\u03B9","\u1F06\u0345","\u1F06\u0399","\u1F06\u03B9","\u1F06\u1FBE","\u1F0E\u0345","\u1F0E\u0399","\u1F0E\u03B9","\u1F0E\u1FBE",}, 160 // \N{GREEK SMALL LETTER ALPHA WITH DASIA AND PERISPOMENI}\N{GREEK SMALL LETTER IOTA} 161 {"\u1F07\u03B9","\u1F07\u0345","\u1F07\u0399","\u1F07\u03B9","\u1F07\u1FBE","\u1F0F\u0345","\u1F0F\u0399","\u1F0F\u03B9","\u1F0F\u1FBE",}, 162 // \N{GREEK SMALL LETTER ETA WITH PSILI}\N{GREEK SMALL LETTER IOTA} 163 {"\u1F20\u03B9","\u1F20\u0345","\u1F20\u0399","\u1F20\u03B9","\u1F20\u1FBE","\u1F28\u0345","\u1F28\u0399","\u1F28\u03B9","\u1F28\u1FBE",}, 164 // \N{GREEK SMALL LETTER ETA WITH DASIA}\N{GREEK SMALL LETTER IOTA} 165 {"\u1F21\u03B9","\u1F21\u0345","\u1F21\u0399","\u1F21\u03B9","\u1F21\u1FBE","\u1F29\u0345","\u1F29\u0399","\u1F29\u03B9","\u1F29\u1FBE",}, 166 // \N{GREEK SMALL LETTER ETA WITH PSILI AND VARIA}\N{GREEK SMALL LETTER IOTA} 167 {"\u1F22\u03B9","\u1F22\u0345","\u1F22\u0399","\u1F22\u03B9","\u1F22\u1FBE","\u1F2A\u0345","\u1F2A\u0399","\u1F2A\u03B9","\u1F2A\u1FBE",}, 168 // \N{GREEK SMALL LETTER ETA WITH DASIA AND VARIA}\N{GREEK SMALL LETTER IOTA} 169 {"\u1F23\u03B9","\u1F23\u0345","\u1F23\u0399","\u1F23\u03B9","\u1F23\u1FBE","\u1F2B\u0345","\u1F2B\u0399","\u1F2B\u03B9","\u1F2B\u1FBE",}, 170 // \N{GREEK SMALL LETTER ETA WITH PSILI AND OXIA}\N{GREEK SMALL LETTER IOTA} 171 {"\u1F24\u03B9","\u1F24\u0345","\u1F24\u0399","\u1F24\u03B9","\u1F24\u1FBE","\u1F2C\u0345","\u1F2C\u0399","\u1F2C\u03B9","\u1F2C\u1FBE",}, 172 // \N{GREEK SMALL LETTER ETA WITH DASIA AND OXIA}\N{GREEK SMALL LETTER IOTA} 173 {"\u1F25\u03B9","\u1F25\u0345","\u1F25\u0399","\u1F25\u03B9","\u1F25\u1FBE","\u1F2D\u0345","\u1F2D\u0399","\u1F2D\u03B9","\u1F2D\u1FBE",}, 174 // \N{GREEK SMALL LETTER ETA WITH PSILI AND PERISPOMENI}\N{GREEK SMALL LETTER IOTA} 175 {"\u1F26\u03B9","\u1F26\u0345","\u1F26\u0399","\u1F26\u03B9","\u1F26\u1FBE","\u1F2E\u0345","\u1F2E\u0399","\u1F2E\u03B9","\u1F2E\u1FBE",}, 176 // \N{GREEK SMALL LETTER ETA WITH DASIA AND PERISPOMENI}\N{GREEK SMALL LETTER IOTA} 177 {"\u1F27\u03B9","\u1F27\u0345","\u1F27\u0399","\u1F27\u03B9","\u1F27\u1FBE","\u1F2F\u0345","\u1F2F\u0399","\u1F2F\u03B9","\u1F2F\u1FBE",}, 178 // \N{GREEK SMALL LETTER OMEGA WITH PSILI}\N{GREEK SMALL LETTER IOTA} 179 {"\u1F60\u03B9","\u1F60\u0345","\u1F60\u0399","\u1F60\u03B9","\u1F60\u1FBE","\u1F68\u0345","\u1F68\u0399","\u1F68\u03B9","\u1F68\u1FBE",}, 180 // \N{GREEK SMALL LETTER OMEGA WITH DASIA}\N{GREEK SMALL LETTER IOTA} 181 {"\u1F61\u03B9","\u1F61\u0345","\u1F61\u0399","\u1F61\u03B9","\u1F61\u1FBE","\u1F69\u0345","\u1F69\u0399","\u1F69\u03B9","\u1F69\u1FBE",}, 182 // \N{GREEK SMALL LETTER OMEGA WITH PSILI AND VARIA}\N{GREEK SMALL LETTER IOTA} 183 {"\u1F62\u03B9","\u1F62\u0345","\u1F62\u0399","\u1F62\u03B9","\u1F62\u1FBE","\u1F6A\u0345","\u1F6A\u0399","\u1F6A\u03B9","\u1F6A\u1FBE",}, 184 // \N{GREEK SMALL LETTER OMEGA WITH DASIA AND VARIA}\N{GREEK SMALL LETTER IOTA} 185 {"\u1F63\u03B9","\u1F63\u0345","\u1F63\u0399","\u1F63\u03B9","\u1F63\u1FBE","\u1F6B\u0345","\u1F6B\u0399","\u1F6B\u03B9","\u1F6B\u1FBE",}, 186 // \N{GREEK SMALL LETTER OMEGA WITH PSILI AND OXIA}\N{GREEK SMALL LETTER IOTA} 187 {"\u1F64\u03B9","\u1F64\u0345","\u1F64\u0399","\u1F64\u03B9","\u1F64\u1FBE","\u1F6C\u0345","\u1F6C\u0399","\u1F6C\u03B9","\u1F6C\u1FBE",}, 188 // \N{GREEK SMALL LETTER OMEGA WITH DASIA AND OXIA}\N{GREEK SMALL LETTER IOTA} 189 {"\u1F65\u03B9","\u1F65\u0345","\u1F65\u0399","\u1F65\u03B9","\u1F65\u1FBE","\u1F6D\u0345","\u1F6D\u0399","\u1F6D\u03B9","\u1F6D\u1FBE",}, 190 // \N{GREEK SMALL LETTER OMEGA WITH PSILI AND PERISPOMENI}\N{GREEK SMALL LETTER IOTA} 191 {"\u1F66\u03B9","\u1F66\u0345","\u1F66\u0399","\u1F66\u03B9","\u1F66\u1FBE","\u1F6E\u0345","\u1F6E\u0399","\u1F6E\u03B9","\u1F6E\u1FBE",}, 192 // \N{GREEK SMALL LETTER OMEGA WITH DASIA AND PERISPOMENI}\N{GREEK SMALL LETTER IOTA} 193 {"\u1F67\u03B9","\u1F67\u0345","\u1F67\u0399","\u1F67\u03B9","\u1F67\u1FBE","\u1F6F\u0345","\u1F6F\u0399","\u1F6F\u03B9","\u1F6F\u1FBE",}, 194 // \N{GREEK SMALL LETTER ALPHA WITH VARIA}\N{GREEK SMALL LETTER IOTA} 195 {"\u1F70\u03B9","\u1F70\u0345","\u1F70\u0399","\u1F70\u03B9","\u1F70\u1FBE","\u1FBA\u0345","\u1FBA\u0399","\u1FBA\u03B9","\u1FBA\u1FBE",}, 196 // \N{GREEK SMALL LETTER ETA WITH VARIA}\N{GREEK SMALL LETTER IOTA} 197 {"\u1F74\u03B9","\u1F74\u0345","\u1F74\u0399","\u1F74\u03B9","\u1F74\u1FBE","\u1FCA\u0345","\u1FCA\u0399","\u1FCA\u03B9","\u1FCA\u1FBE",}, 198 // \N{GREEK SMALL LETTER OMEGA WITH VARIA}\N{GREEK SMALL LETTER IOTA} 199 {"\u1F7C\u03B9","\u1F7C\u0345","\u1F7C\u0399","\u1F7C\u03B9","\u1F7C\u1FBE","\u1FFA\u0345","\u1FFA\u0399","\u1FFA\u03B9","\u1FFA\u1FBE",}, 200 }; 201 202 // this initializes the data used to generated the case-equivalents 203 204 static { 205 206 // Gather up the exceptions in a form we can use 207 208 if (!GENERATE) { 209 for (int i = 0; i < exceptionList.length; ++i) { 210 String[] exception = exceptionList[i]; 211 Set s = new HashSet(); 212 // there has to be some method to do the following, but I can't find it in the collections 213 for (int j = 0; j < exception.length; ++j) { 214 s.add(exception[j]); 215 } 216 fromCaseFold.put(exception[0], s); 217 } 218 } 219 220 // walk through all the characters, and at every case fold result, 221 // put a set of all the characters that map to that result 222 223 boolean defaultmapping = true; // false for turkish 224 for (int i = 0; i <= 0x10FFFF; ++i) { 225 int cat = UCharacter.getType(i); 226 if (cat == Character.UNASSIGNED || cat == Character.PRIVATE_USE) continue; 227 228 String cp = UTF16.valueOf(i); 229 String mapped = UCharacter.foldCase(cp, defaultmapping); 230 if (mapped.equals(cp)) continue; 231 232 if (maxLength < mapped.length()) maxLength = mapped.length(); 233 234 // at this point, have different case folding 235 236 Set s = (Set) fromCaseFold.get(mapped); 237 if (s == null) { 238 s = new HashSet(); 239 s.add(mapped); // add the case fold result itself 240 fromCaseFold.put(mapped, s); 241 } 242 s.add(cp); 243 toCaseFold.put(cp, mapped); 244 toCaseFold.put(mapped, mapped); // add mapping to self 245 } 246 247 // Emit the final data 248 249 if (DUMP) { 250 System.out.println("maxLength = " + maxLength); 251 252 System.out.println("\nfromCaseFold:"); 253 Iterator it = fromCaseFold.keySet().iterator(); 254 while (it.hasNext()) { 255 Object key = it.next(); 256 System.out.print(" " + toHex2.transliterate((String)key) + ": "); 257 Set s = (Set) fromCaseFold.get(key); 258 Iterator it2 = s.iterator(); 259 boolean first = true; 260 while (it2.hasNext()) { 261 if (first) { 262 first = false; 263 } else { 264 System.out.print(", "); 265 } 266 System.out.print(toHex2.transliterate((String)it2.next())); 267 } 268 System.out.println(""); 269 } 270 271 System.out.println("\ntoCaseFold:"); 272 it = toCaseFold.keySet().iterator(); 273 while (it.hasNext()) { 274 String key = (String) it.next(); 275 String value = (String) toCaseFold.get(key); 276 System.out.println(" " + toHex2.transliterate(key) + ": " + toHex2.transliterate(value)); 277 } 278 } 279 280 // Now convert all those sets into linear arrays 281 // We can't do this in place in Java, so make a temporary target array 282 283 // Note: This could be transformed into a single array, with offsets into it. 284 // Might be best choice in C. 285 286 287 Map fromCaseFold2 = new HashMap(); 288 Iterator it = fromCaseFold.keySet().iterator(); 289 while (it.hasNext()) { 290 Object key = it.next(); 291 Set s = (Set) fromCaseFold.get(key); 292 String[] temp = new String[s.size()]; 293 s.toArray(temp); 294 fromCaseFold2.put(key, temp); 295 } 296 fromCaseFold = fromCaseFold2; 297 298 // We have processed everything, so the iterator will now work 299 // The following is normally OFF. 300 // It is here to generate (under the GENERATE flag) the static exception list. 301 // It must be at the very end of initialization, so that the iterator is functional. 302 // (easiest to do it that way) 303 304 if (GENERATE) { 305 306 // first get small set of items that have multiple characters 307 308 Set multichars = new TreeSet(); 309 it = fromCaseFold.keySet().iterator(); 310 while (it.hasNext()) { 311 String key = (String) it.next(); 312 if (UTF16.countCodePoint(key) < 2) continue; 313 multichars.add(key); 314 } 315 316 // now we will go through each of them. 317 318 CaseIterator ci = new CaseIterator(); 319 it = multichars.iterator(); 320 321 while (it.hasNext()) { 322 String key = (String) it.next(); 323 324 // here is a nasty complication. Take 'ffi' ligature. We 325 // can't just close it, since we would miss the combination 326 // that includes the 'fi' => "fi" ligature 327 // so first do a pass through, and add substring combinations 328 // we call this a 'partial closure' 329 330 Set partialClosure = new TreeSet(); 331 partialClosure.add(key); 332 333 if (UTF16.countCodePoint(key) > 2) { 334 Iterator multiIt2 = multichars.iterator(); 335 while (multiIt2.hasNext()) { 336 String otherKey = (String) multiIt2.next(); 337 if (otherKey.length() >= key.length()) continue; 338 int pos = -1; 339 while (true) { 340 // The following is not completely general 341 // but works for the actual cased stuff, 342 // and should work for future characters, since we won't have 343 // more ligatures & other oddities. 344 pos = key.indexOf(otherKey, pos+1); 345 if (pos < 0) break; 346 int endPos = pos + otherKey.length(); 347 // we know we have a proper substring, 348 // so get the combinations 349 String[] choices = (String[]) fromCaseFold.get(otherKey); 350 for (int ii = 0; ii < choices.length; ++ii) { 351 String patchwork = key.substring(0, pos) 352 + choices[ii] 353 + key.substring(endPos); 354 partialClosure.add(patchwork); 355 } 356 } 357 } 358 } 359 360 // now, for each thing in the partial closure, get its 361 // case closure and add it to the final result. 362 363 Set closure = new TreeSet(); // this will be the real closure 364 Iterator partialIt = partialClosure.iterator(); 365 while (partialIt.hasNext()) { 366 String key2 = (String) partialIt.next(); 367 ci.reset(key2); 368 for (String temp = ci.next(); temp != null; temp = ci.next()) { 369 closure.add(temp); 370 } 371 // form closure 372 /*String[] choices = (String[]) fromCaseFold.get(key2); 373 for (int i = 0; i < choices.length; ++i) { 374 ci.reset(choices[i]); 375 String temp; 376 while (null != (temp = ci.next())) { 377 closure.add(temp); 378 } 379 } 380 */ 381 } 382 383 // print it out, so that it can be cut and pasted back into this document. 384 385 Iterator it2 = closure.iterator(); 386 System.out.println("\t// " + toName.transliterate(key)); 387 System.out.print("\t{\"" + toHex.transliterate(key) + "\","); 388 while (it2.hasNext()) { 389 String item = (String)it2.next(); 390 System.out.print("\"" + toHex.transliterate(item) + "\","); 391 } 392 System.out.println("},"); 393 } 394 } 395 } 396 397 // ============ PRIVATE CLASS DATA ============ 398 399 // pieces that we will put together 400 // is not changed during iteration 401 private int count = 0; 402 private String[][] variants; 403 404 // state information, changes during iteration 405 private boolean done = false; 406 private int[] counts; 407 408 // internal buffer for efficiency 409 private StringBuffer nextBuffer = new StringBuffer(); 410 411 // ======================== 412 413 /** 414 * Reset to different source. Once reset, the iteration starts from the beginning. 415 * @param source The string to get case variants for 416 */ 417 public void reset(String source) { 418 419 // allocate arrays to store pieces 420 // using length might be slightly too long, but we don't care much 421 422 counts = new int[source.length()]; 423 variants = new String[source.length()][]; 424 425 // walk through the source, and break up into pieces 426 // each piece becomes an array of equivalent values 427 // TODO: could optimized this later to coalesce all single string pieces 428 429 String piece = null; 430 count = 0; 431 for (int i = 0; i < source.length(); i += piece.length()) { 432 433 // find *longest* matching piece 434 String caseFold = null; 435 436 if (GENERATE) { 437 // do exactly one CP 438 piece = UTF16.valueOf(source, i); 439 caseFold = (String) toCaseFold.get(piece); 440 } else { 441 int max = i + maxLength; 442 if (max > source.length()) max = source.length(); 443 for (int j = max; j > i; --j) { 444 piece = source.substring(i, j); 445 caseFold = (String) toCaseFold.get(piece); 446 if (caseFold != null) break; 447 } 448 } 449 450 // if we fail, pick one code point 451 if (caseFold == null) { 452 piece = UTF16.valueOf(source, i); 453 variants[count++] = new String[] {piece}; // single item string 454 } else { 455 variants[count++] = (String[])fromCaseFold.get(caseFold); 456 } 457 } 458 reset(); 459 } 460 461 /** 462 * Restart the iteration from the beginning, but with same source 463 */ 464 public void reset() { 465 done = false; 466 for (int i = 0; i < count; ++i) { 467 counts[i] = 0; 468 } 469 } 470 471 /** 472 * Iterates through the case variants. 473 * @return next case variant. Each variant will case-fold to the same value as the source will. 474 * When the iteration is done, null is returned. 475 */ 476 public String next() { 477 478 if (done) return null; 479 int i; 480 481 // TODO Optimize so we keep the piece before and after the current position 482 // so we don't have so much concatenation 483 484 // get the result, a concatenation 485 486 nextBuffer.setLength(0); 487 for (i = 0; i < count; ++i) { 488 nextBuffer.append(variants[i][counts[i]]); 489 } 490 491 // find the next right set of pieces to concatenate 492 493 for (i = count-1; i >= 0; --i) { 494 counts[i]++; 495 if (counts[i] < variants[i].length) break; 496 counts[i] = 0; 497 } 498 499 // if we go too far, bail 500 501 if (i < 0) { 502 done = true; 503 } 504 505 return nextBuffer.toString(); 506 } 507 508 509 /** 510 * Temporary test, just to see how the stuff works. 511 */ 512 static public void main(String[] args) { 513 String[] testCases = {"fiss", "h\u03a3"}; 514 CaseIterator ci = new CaseIterator(); 515 516 for (int i = 0; i < testCases.length; ++i) { 517 String item = testCases[i]; 518 System.out.println(); 519 System.out.println("Testing: " + toName.transliterate(item)); 520 System.out.println(); 521 ci.reset(item); 522 int count = 0; 523 for (String temp = ci.next(); temp != null; temp = ci.next()) { 524 System.out.println(toName.transliterate(temp)); 525 count++; 526 } 527 System.out.println("Total: " + count); 528 } 529 530 // generate a list of all caseless characters -- characters whose 531 // case closure is themselves. 532 533 UnicodeSet caseless = new UnicodeSet(); 534 535 for (int i = 0; i <= 0x10FFFF; ++i) { 536 String cp = UTF16.valueOf(i); 537 ci.reset(cp); 538 int count = 0; 539 String fold = null; 540 for (String temp = ci.next(); temp != null; temp = ci.next()) { 541 fold = temp; 542 if (++count > 1) break; 543 } 544 if (count==1 && fold.equals(cp)) { 545 caseless.add(i); 546 } 547 } 548 549 System.out.println("caseless = " + caseless.toPattern(true)); 550 551 UnicodeSet not_lc = new UnicodeSet("[:^lc:]"); 552 553 UnicodeSet a = new UnicodeSet(); 554 a.set(not_lc); 555 a.removeAll(caseless); 556 System.out.println("[:^lc:] - caseless = " + a.toPattern(true)); 557 558 a.set(caseless); 559 a.removeAll(not_lc); 560 System.out.println("caseless - [:^lc:] = " + a.toPattern(true)); 561 } 562} 563