jars/icu4j-4_2_1-src/src/com/ibm/icu/dev/test/util/CaseIterator.java

   1 /**\r
   2 *******************************************************************************\r
   3 * Copyright (C) 1996-2007, International Business Machines Corporation and    *\r
   4 * others. All Rights Reserved.                                                *\r
   5 *******************************************************************************\r
   6 */\r
   7 \r
   8 // copied from the Transliterator demo\r
   9 \r
  10 package com.ibm.icu.dev.test.util;\r
  11 import java.util.*;\r
  12 import com.ibm.icu.lang.UCharacter;\r
  13 import com.ibm.icu.text.UTF16;\r
  14 import com.ibm.icu.text.Transliterator;\r
  15 import com.ibm.icu.text.UnicodeSet;\r
  16 \r
  17 /**\r
  18  * Incrementally returns the set of all strings that case-fold to the same value.\r
  19  */\r
  20 public class CaseIterator {\r
  21     \r
  22     // testing stuff\r
  23     private static Transliterator toName = Transliterator.getInstance("[:^ascii:] Any-Name");\r
  24     private static Transliterator toHex = Transliterator.getInstance("[:^ascii:] Any-Hex");\r
  25     private static Transliterator toHex2 = Transliterator.getInstance("[[^\u0021-\u007F]-[,]] Any-Hex");\r
  26     \r
  27     // global tables (could be precompiled)\r
  28     private static Map fromCaseFold = new HashMap();\r
  29     private static Map toCaseFold = new HashMap();\r
  30     private static int maxLength = 0;\r
  31     \r
  32     // This exception list is generated on the console by turning on the GENERATED flag, \r
  33     // which MUST be false for normal operation.\r
  34     // Once the list is generated, it is pasted in here.\r
  35     // A bit of a cludge, but this bootstrapping is the easiest way \r
  36     // to get around certain complications in the data.\r
  37     \r
  38     private static final boolean GENERATE = false;\r
  39 \r
  40     private static final boolean DUMP = false;\r
  41     \r
  42     private static String[][] exceptionList = {\r
  43         // a\N{MODIFIER LETTER RIGHT HALF RING}\r
  44         {"a\u02BE","A\u02BE","a\u02BE",},\r
  45         // ff\r
  46         {"ff","FF","Ff","fF","ff",},\r
  47         // ffi\r
  48         {"ffi","FFI","FFi","FfI","Ffi","F\uFB01","fFI","fFi","ffI","ffi","f\uFB01","\uFB00I","\uFB00i",},\r
  49         // ffl\r
  50         {"ffl","FFL","FFl","FfL","Ffl","F\uFB02","fFL","fFl","ffL","ffl","f\uFB02","\uFB00L","\uFB00l",},\r
  51         // fi\r
  52         {"fi","FI","Fi","fI","fi",},\r
  53         // fl\r
  54         {"fl","FL","Fl","fL","fl",},\r
  55         // h\N{COMBINING MACRON BELOW}\r
  56         {"h\u0331","H\u0331","h\u0331",},\r
  57         // i\N{COMBINING DOT ABOVE}\r
  58         {"i\u0307","I\u0307","i\u0307",},\r
  59         // j\N{COMBINING CARON}\r
  60         {"j\u030C","J\u030C","j\u030C",},\r
  61         // ss\r
  62         {"ss","SS","Ss","S\u017F","sS","ss","s\u017F","\u017FS","\u017Fs","\u017F\u017F",},\r
  63         // st\r
  64         {"st","ST","St","sT","st","\u017FT","\u017Ft",},\r
  65         // t\N{COMBINING DIAERESIS}\r
  66         {"t\u0308","T\u0308","t\u0308",},\r
  67         // w\N{COMBINING RING ABOVE}\r
  68         {"w\u030A","W\u030A","w\u030A",},\r
  69         // y\N{COMBINING RING ABOVE}\r
  70         {"y\u030A","Y\u030A","y\u030A",},\r
  71         // \N{MODIFIER LETTER APOSTROPHE}n\r
  72         {"\u02BCn","\u02BCN","\u02BCn",},\r
  73         // \N{GREEK SMALL LETTER ALPHA WITH TONOS}\N{GREEK SMALL LETTER IOTA}\r
  74         {"\u03AC\u03B9","\u0386\u0345","\u0386\u0399","\u0386\u03B9","\u0386\u1FBE","\u03AC\u0345","\u03AC\u0399","\u03AC\u03B9","\u03AC\u1FBE",},\r
  75         // \N{GREEK SMALL LETTER ETA WITH TONOS}\N{GREEK SMALL LETTER IOTA}\r
  76         {"\u03AE\u03B9","\u0389\u0345","\u0389\u0399","\u0389\u03B9","\u0389\u1FBE","\u03AE\u0345","\u03AE\u0399","\u03AE\u03B9","\u03AE\u1FBE",},\r
  77         // \N{GREEK SMALL LETTER ALPHA}\N{COMBINING GREEK PERISPOMENI}\r
  78         {"\u03B1\u0342","\u0391\u0342","\u03B1\u0342",},\r
  79         // \N{GREEK SMALL LETTER ALPHA}\N{COMBINING GREEK PERISPOMENI}\N{GREEK SMALL LETTER IOTA}\r
  80         {"\u03B1\u0342\u03B9","\u0391\u0342\u0345","\u0391\u0342\u0399","\u0391\u0342\u03B9","\u0391\u0342\u1FBE",\r
  81             "\u03B1\u0342\u0345","\u03B1\u0342\u0399","\u03B1\u0342\u03B9","\u03B1\u0342\u1FBE","\u1FB6\u0345",\r
  82             "\u1FB6\u0399","\u1FB6\u03B9","\u1FB6\u1FBE",},\r
  83         // \N{GREEK SMALL LETTER ALPHA}\N{GREEK SMALL LETTER IOTA}\r
  84         {"\u03B1\u03B9","\u0391\u0345","\u0391\u0399","\u0391\u03B9","\u0391\u1FBE","\u03B1\u0345","\u03B1\u0399","\u03B1\u03B9","\u03B1\u1FBE",},\r
  85         // \N{GREEK SMALL LETTER ETA}\N{COMBINING GREEK PERISPOMENI}\r
  86         {"\u03B7\u0342","\u0397\u0342","\u03B7\u0342",},\r
  87         // \N{GREEK SMALL LETTER ETA}\N{COMBINING GREEK PERISPOMENI}\N{GREEK SMALL LETTER IOTA}\r
  88         {"\u03B7\u0342\u03B9","\u0397\u0342\u0345","\u0397\u0342\u0399","\u0397\u0342\u03B9","\u0397\u0342\u1FBE",\r
  89             "\u03B7\u0342\u0345","\u03B7\u0342\u0399","\u03B7\u0342\u03B9","\u03B7\u0342\u1FBE","\u1FC6\u0345","\u1FC6\u0399",\r
  90             "\u1FC6\u03B9","\u1FC6\u1FBE",},\r
  91         // \N{GREEK SMALL LETTER ETA}\N{GREEK SMALL LETTER IOTA}\r
  92         {"\u03B7\u03B9","\u0397\u0345","\u0397\u0399","\u0397\u03B9","\u0397\u1FBE","\u03B7\u0345","\u03B7\u0399","\u03B7\u03B9","\u03B7\u1FBE",},\r
  93         // \N{GREEK SMALL LETTER IOTA}\N{COMBINING DIAERESIS}\N{COMBINING GRAVE ACCENT}\r
  94         {"\u03B9\u0308\u0300","\u0345\u0308\u0300","\u0399\u0308\u0300","\u03B9\u0308\u0300","\u1FBE\u0308\u0300",},\r
  95         // \N{GREEK SMALL LETTER IOTA}\N{COMBINING DIAERESIS}\N{COMBINING ACUTE ACCENT}\r
  96         {"\u03B9\u0308\u0301","\u0345\u0308\u0301","\u0399\u0308\u0301","\u03B9\u0308\u0301","\u1FBE\u0308\u0301",},\r
  97         // \N{GREEK SMALL LETTER IOTA}\N{COMBINING DIAERESIS}\N{COMBINING GREEK PERISPOMENI}\r
  98         {"\u03B9\u0308\u0342","\u0345\u0308\u0342","\u0399\u0308\u0342","\u03B9\u0308\u0342","\u1FBE\u0308\u0342",},\r
  99         // \N{GREEK SMALL LETTER IOTA}\N{COMBINING GREEK PERISPOMENI}\r
 100         {"\u03B9\u0342","\u0345\u0342","\u0399\u0342","\u03B9\u0342","\u1FBE\u0342",},\r
 101         // \N{GREEK SMALL LETTER RHO}\N{COMBINING COMMA ABOVE}\r
 102         {"\u03C1\u0313","\u03A1\u0313","\u03C1\u0313","\u03F1\u0313",},\r
 103         // \N{GREEK SMALL LETTER UPSILON}\N{COMBINING DIAERESIS}\N{COMBINING GRAVE ACCENT}\r
 104         {"\u03C5\u0308\u0300","\u03A5\u0308\u0300","\u03C5\u0308\u0300",},\r
 105         // \N{GREEK SMALL LETTER UPSILON}\N{COMBINING DIAERESIS}\N{COMBINING ACUTE ACCENT}\r
 106         {"\u03C5\u0308\u0301","\u03A5\u0308\u0301","\u03C5\u0308\u0301",},\r
 107         // \N{GREEK SMALL LETTER UPSILON}\N{COMBINING DIAERESIS}\N{COMBINING GREEK PERISPOMENI}\r
 108         {"\u03C5\u0308\u0342","\u03A5\u0308\u0342","\u03C5\u0308\u0342",},\r
 109         // \N{GREEK SMALL LETTER UPSILON}\N{COMBINING COMMA ABOVE}\r
 110         {"\u03C5\u0313","\u03A5\u0313","\u03C5\u0313",},\r
 111         // \N{GREEK SMALL LETTER UPSILON}\N{COMBINING COMMA ABOVE}\N{COMBINING GRAVE ACCENT}\r
 112         {"\u03C5\u0313\u0300","\u03A5\u0313\u0300","\u03C5\u0313\u0300","\u1F50\u0300",},\r
 113         // \N{GREEK SMALL LETTER UPSILON}\N{COMBINING COMMA ABOVE}\N{COMBINING ACUTE ACCENT}\r
 114         {"\u03C5\u0313\u0301","\u03A5\u0313\u0301","\u03C5\u0313\u0301","\u1F50\u0301",},\r
 115         // \N{GREEK SMALL LETTER UPSILON}\N{COMBINING COMMA ABOVE}\N{COMBINING GREEK PERISPOMENI}\r
 116         {"\u03C5\u0313\u0342","\u03A5\u0313\u0342","\u03C5\u0313\u0342","\u1F50\u0342",},\r
 117         // \N{GREEK SMALL LETTER UPSILON}\N{COMBINING GREEK PERISPOMENI}\r
 118         {"\u03C5\u0342","\u03A5\u0342","\u03C5\u0342",},\r
 119         // \N{GREEK SMALL LETTER OMEGA}\N{COMBINING GREEK PERISPOMENI}\r
 120         {"\u03C9\u0342","\u03A9\u0342","\u03C9\u0342","\u2126\u0342",},\r
 121         // \N{GREEK SMALL LETTER OMEGA}\N{COMBINING GREEK PERISPOMENI}\N{GREEK SMALL LETTER IOTA}\r
 122         {"\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",\r
 123             "\u1FF6\u0399","\u1FF6\u03B9","\u1FF6\u1FBE","\u2126\u0342\u0345","\u2126\u0342\u0399","\u2126\u0342\u03B9","\u2126\u0342\u1FBE",},\r
 124         // \N{GREEK SMALL LETTER OMEGA}\N{GREEK SMALL LETTER IOTA}\r
 125         {"\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",},\r
 126         // \N{GREEK SMALL LETTER OMEGA WITH TONOS}\N{GREEK SMALL LETTER IOTA}\r
 127         {"\u03CE\u03B9","\u038F\u0345","\u038F\u0399","\u038F\u03B9","\u038F\u1FBE","\u03CE\u0345","\u03CE\u0399","\u03CE\u03B9","\u03CE\u1FBE",},\r
 128         // \N{ARMENIAN SMALL LETTER ECH}\N{ARMENIAN SMALL LETTER YIWN}\r
 129         {"\u0565\u0582","\u0535\u0552","\u0535\u0582","\u0565\u0552","\u0565\u0582",},\r
 130         // \N{ARMENIAN SMALL LETTER MEN}\N{ARMENIAN SMALL LETTER ECH}\r
 131         {"\u0574\u0565","\u0544\u0535","\u0544\u0565","\u0574\u0535","\u0574\u0565",},\r
 132         // \N{ARMENIAN SMALL LETTER MEN}\N{ARMENIAN SMALL LETTER INI}\r
 133         {"\u0574\u056B","\u0544\u053B","\u0544\u056B","\u0574\u053B","\u0574\u056B",},\r
 134         // \N{ARMENIAN SMALL LETTER MEN}\N{ARMENIAN SMALL LETTER XEH}\r
 135         {"\u0574\u056D","\u0544\u053D","\u0544\u056D","\u0574\u053D","\u0574\u056D",},\r
 136         // \N{ARMENIAN SMALL LETTER MEN}\N{ARMENIAN SMALL LETTER NOW}\r
 137         {"\u0574\u0576","\u0544\u0546","\u0544\u0576","\u0574\u0546","\u0574\u0576",},\r
 138         // \N{ARMENIAN SMALL LETTER VEW}\N{ARMENIAN SMALL LETTER NOW}\r
 139         {"\u057E\u0576","\u054E\u0546","\u054E\u0576","\u057E\u0546","\u057E\u0576",},\r
 140         // \N{GREEK SMALL LETTER ALPHA WITH PSILI}\N{GREEK SMALL LETTER IOTA}\r
 141         {"\u1F00\u03B9","\u1F00\u0345","\u1F00\u0399","\u1F00\u03B9","\u1F00\u1FBE","\u1F08\u0345","\u1F08\u0399","\u1F08\u03B9","\u1F08\u1FBE",},\r
 142         // \N{GREEK SMALL LETTER ALPHA WITH DASIA}\N{GREEK SMALL LETTER IOTA}\r
 143         {"\u1F01\u03B9","\u1F01\u0345","\u1F01\u0399","\u1F01\u03B9","\u1F01\u1FBE","\u1F09\u0345","\u1F09\u0399","\u1F09\u03B9","\u1F09\u1FBE",},\r
 144         // \N{GREEK SMALL LETTER ALPHA WITH PSILI AND VARIA}\N{GREEK SMALL LETTER IOTA}\r
 145         {"\u1F02\u03B9","\u1F02\u0345","\u1F02\u0399","\u1F02\u03B9","\u1F02\u1FBE","\u1F0A\u0345","\u1F0A\u0399","\u1F0A\u03B9","\u1F0A\u1FBE",},\r
 146         // \N{GREEK SMALL LETTER ALPHA WITH DASIA AND VARIA}\N{GREEK SMALL LETTER IOTA}\r
 147         {"\u1F03\u03B9","\u1F03\u0345","\u1F03\u0399","\u1F03\u03B9","\u1F03\u1FBE","\u1F0B\u0345","\u1F0B\u0399","\u1F0B\u03B9","\u1F0B\u1FBE",},\r
 148         // \N{GREEK SMALL LETTER ALPHA WITH PSILI AND OXIA}\N{GREEK SMALL LETTER IOTA}\r
 149         {"\u1F04\u03B9","\u1F04\u0345","\u1F04\u0399","\u1F04\u03B9","\u1F04\u1FBE","\u1F0C\u0345","\u1F0C\u0399","\u1F0C\u03B9","\u1F0C\u1FBE",},\r
 150         // \N{GREEK SMALL LETTER ALPHA WITH DASIA AND OXIA}\N{GREEK SMALL LETTER IOTA}\r
 151         {"\u1F05\u03B9","\u1F05\u0345","\u1F05\u0399","\u1F05\u03B9","\u1F05\u1FBE","\u1F0D\u0345","\u1F0D\u0399","\u1F0D\u03B9","\u1F0D\u1FBE",},\r
 152         // \N{GREEK SMALL LETTER ALPHA WITH PSILI AND PERISPOMENI}\N{GREEK SMALL LETTER IOTA}\r
 153         {"\u1F06\u03B9","\u1F06\u0345","\u1F06\u0399","\u1F06\u03B9","\u1F06\u1FBE","\u1F0E\u0345","\u1F0E\u0399","\u1F0E\u03B9","\u1F0E\u1FBE",},\r
 154         // \N{GREEK SMALL LETTER ALPHA WITH DASIA AND PERISPOMENI}\N{GREEK SMALL LETTER IOTA}\r
 155         {"\u1F07\u03B9","\u1F07\u0345","\u1F07\u0399","\u1F07\u03B9","\u1F07\u1FBE","\u1F0F\u0345","\u1F0F\u0399","\u1F0F\u03B9","\u1F0F\u1FBE",},\r
 156         // \N{GREEK SMALL LETTER ETA WITH PSILI}\N{GREEK SMALL LETTER IOTA}\r
 157         {"\u1F20\u03B9","\u1F20\u0345","\u1F20\u0399","\u1F20\u03B9","\u1F20\u1FBE","\u1F28\u0345","\u1F28\u0399","\u1F28\u03B9","\u1F28\u1FBE",},\r
 158         // \N{GREEK SMALL LETTER ETA WITH DASIA}\N{GREEK SMALL LETTER IOTA}\r
 159         {"\u1F21\u03B9","\u1F21\u0345","\u1F21\u0399","\u1F21\u03B9","\u1F21\u1FBE","\u1F29\u0345","\u1F29\u0399","\u1F29\u03B9","\u1F29\u1FBE",},\r
 160         // \N{GREEK SMALL LETTER ETA WITH PSILI AND VARIA}\N{GREEK SMALL LETTER IOTA}\r
 161         {"\u1F22\u03B9","\u1F22\u0345","\u1F22\u0399","\u1F22\u03B9","\u1F22\u1FBE","\u1F2A\u0345","\u1F2A\u0399","\u1F2A\u03B9","\u1F2A\u1FBE",},\r
 162         // \N{GREEK SMALL LETTER ETA WITH DASIA AND VARIA}\N{GREEK SMALL LETTER IOTA}\r
 163         {"\u1F23\u03B9","\u1F23\u0345","\u1F23\u0399","\u1F23\u03B9","\u1F23\u1FBE","\u1F2B\u0345","\u1F2B\u0399","\u1F2B\u03B9","\u1F2B\u1FBE",},\r
 164         // \N{GREEK SMALL LETTER ETA WITH PSILI AND OXIA}\N{GREEK SMALL LETTER IOTA}\r
 165         {"\u1F24\u03B9","\u1F24\u0345","\u1F24\u0399","\u1F24\u03B9","\u1F24\u1FBE","\u1F2C\u0345","\u1F2C\u0399","\u1F2C\u03B9","\u1F2C\u1FBE",},\r
 166         // \N{GREEK SMALL LETTER ETA WITH DASIA AND OXIA}\N{GREEK SMALL LETTER IOTA}\r
 167         {"\u1F25\u03B9","\u1F25\u0345","\u1F25\u0399","\u1F25\u03B9","\u1F25\u1FBE","\u1F2D\u0345","\u1F2D\u0399","\u1F2D\u03B9","\u1F2D\u1FBE",},\r
 168         // \N{GREEK SMALL LETTER ETA WITH PSILI AND PERISPOMENI}\N{GREEK SMALL LETTER IOTA}\r
 169         {"\u1F26\u03B9","\u1F26\u0345","\u1F26\u0399","\u1F26\u03B9","\u1F26\u1FBE","\u1F2E\u0345","\u1F2E\u0399","\u1F2E\u03B9","\u1F2E\u1FBE",},\r
 170         // \N{GREEK SMALL LETTER ETA WITH DASIA AND PERISPOMENI}\N{GREEK SMALL LETTER IOTA}\r
 171         {"\u1F27\u03B9","\u1F27\u0345","\u1F27\u0399","\u1F27\u03B9","\u1F27\u1FBE","\u1F2F\u0345","\u1F2F\u0399","\u1F2F\u03B9","\u1F2F\u1FBE",},\r
 172         // \N{GREEK SMALL LETTER OMEGA WITH PSILI}\N{GREEK SMALL LETTER IOTA}\r
 173         {"\u1F60\u03B9","\u1F60\u0345","\u1F60\u0399","\u1F60\u03B9","\u1F60\u1FBE","\u1F68\u0345","\u1F68\u0399","\u1F68\u03B9","\u1F68\u1FBE",},\r
 174         // \N{GREEK SMALL LETTER OMEGA WITH DASIA}\N{GREEK SMALL LETTER IOTA}\r
 175         {"\u1F61\u03B9","\u1F61\u0345","\u1F61\u0399","\u1F61\u03B9","\u1F61\u1FBE","\u1F69\u0345","\u1F69\u0399","\u1F69\u03B9","\u1F69\u1FBE",},\r
 176         // \N{GREEK SMALL LETTER OMEGA WITH PSILI AND VARIA}\N{GREEK SMALL LETTER IOTA}\r
 177         {"\u1F62\u03B9","\u1F62\u0345","\u1F62\u0399","\u1F62\u03B9","\u1F62\u1FBE","\u1F6A\u0345","\u1F6A\u0399","\u1F6A\u03B9","\u1F6A\u1FBE",},\r
 178         // \N{GREEK SMALL LETTER OMEGA WITH DASIA AND VARIA}\N{GREEK SMALL LETTER IOTA}\r
 179         {"\u1F63\u03B9","\u1F63\u0345","\u1F63\u0399","\u1F63\u03B9","\u1F63\u1FBE","\u1F6B\u0345","\u1F6B\u0399","\u1F6B\u03B9","\u1F6B\u1FBE",},\r
 180         // \N{GREEK SMALL LETTER OMEGA WITH PSILI AND OXIA}\N{GREEK SMALL LETTER IOTA}\r
 181         {"\u1F64\u03B9","\u1F64\u0345","\u1F64\u0399","\u1F64\u03B9","\u1F64\u1FBE","\u1F6C\u0345","\u1F6C\u0399","\u1F6C\u03B9","\u1F6C\u1FBE",},\r
 182         // \N{GREEK SMALL LETTER OMEGA WITH DASIA AND OXIA}\N{GREEK SMALL LETTER IOTA}\r
 183         {"\u1F65\u03B9","\u1F65\u0345","\u1F65\u0399","\u1F65\u03B9","\u1F65\u1FBE","\u1F6D\u0345","\u1F6D\u0399","\u1F6D\u03B9","\u1F6D\u1FBE",},\r
 184         // \N{GREEK SMALL LETTER OMEGA WITH PSILI AND PERISPOMENI}\N{GREEK SMALL LETTER IOTA}\r
 185         {"\u1F66\u03B9","\u1F66\u0345","\u1F66\u0399","\u1F66\u03B9","\u1F66\u1FBE","\u1F6E\u0345","\u1F6E\u0399","\u1F6E\u03B9","\u1F6E\u1FBE",},\r
 186         // \N{GREEK SMALL LETTER OMEGA WITH DASIA AND PERISPOMENI}\N{GREEK SMALL LETTER IOTA}\r
 187         {"\u1F67\u03B9","\u1F67\u0345","\u1F67\u0399","\u1F67\u03B9","\u1F67\u1FBE","\u1F6F\u0345","\u1F6F\u0399","\u1F6F\u03B9","\u1F6F\u1FBE",},\r
 188         // \N{GREEK SMALL LETTER ALPHA WITH VARIA}\N{GREEK SMALL LETTER IOTA}\r
 189         {"\u1F70\u03B9","\u1F70\u0345","\u1F70\u0399","\u1F70\u03B9","\u1F70\u1FBE","\u1FBA\u0345","\u1FBA\u0399","\u1FBA\u03B9","\u1FBA\u1FBE",},\r
 190         // \N{GREEK SMALL LETTER ETA WITH VARIA}\N{GREEK SMALL LETTER IOTA}\r
 191         {"\u1F74\u03B9","\u1F74\u0345","\u1F74\u0399","\u1F74\u03B9","\u1F74\u1FBE","\u1FCA\u0345","\u1FCA\u0399","\u1FCA\u03B9","\u1FCA\u1FBE",},\r
 192         // \N{GREEK SMALL LETTER OMEGA WITH VARIA}\N{GREEK SMALL LETTER IOTA}\r
 193         {"\u1F7C\u03B9","\u1F7C\u0345","\u1F7C\u0399","\u1F7C\u03B9","\u1F7C\u1FBE","\u1FFA\u0345","\u1FFA\u0399","\u1FFA\u03B9","\u1FFA\u1FBE",},\r
 194     };\r
 195     \r
 196     // this initializes the data used to generated the case-equivalents\r
 197 \r
 198     static {\r
 199         \r
 200         // Gather up the exceptions in a form we can use\r
 201         \r
 202         if (!GENERATE) {\r
 203             for (int i = 0; i < exceptionList.length; ++i) {\r
 204                 String[] exception = exceptionList[i];\r
 205                 Set s = new HashSet();\r
 206                 // there has to be some method to do the following, but I can't find it in the collections\r
 207                 for (int j = 0; j < exception.length; ++j) {\r
 208                     s.add(exception[j]);\r
 209                 }\r
 210                 fromCaseFold.put(exception[0], s);\r
 211             }\r
 212         }\r
 213         \r
 214         // walk through all the characters, and at every case fold result,\r
 215         // put a set of all the characters that map to that result\r
 216 \r
 217         boolean defaultmapping = true; // false for turkish\r
 218         for (int i = 0; i <= 0x10FFFF; ++i) {\r
 219             int cat = UCharacter.getType(i);\r
 220             if (cat == Character.UNASSIGNED || cat == Character.PRIVATE_USE) continue;\r
 221             \r
 222             String cp = UTF16.valueOf(i);\r
 223             String mapped = UCharacter.foldCase(cp, defaultmapping);\r
 224             if (mapped.equals(cp)) continue;\r
 225             \r
 226             if (maxLength < mapped.length()) maxLength = mapped.length();\r
 227             \r
 228             // at this point, have different case folding\r
 229             \r
 230             Set s = (Set) fromCaseFold.get(mapped);\r
 231             if (s == null) {\r
 232                 s = new HashSet();\r
 233                 s.add(mapped); // add the case fold result itself\r
 234                 fromCaseFold.put(mapped, s);\r
 235             }\r
 236             s.add(cp);\r
 237             toCaseFold.put(cp, mapped);\r
 238             toCaseFold.put(mapped, mapped); // add mapping to self\r
 239         }\r
 240         \r
 241         // Emit the final data\r
 242 \r
 243         if (DUMP) {\r
 244             System.out.println("maxLength = " + maxLength);\r
 245 \r
 246             System.out.println("\nfromCaseFold:");\r
 247             Iterator it = fromCaseFold.keySet().iterator();\r
 248             while (it.hasNext()) {\r
 249                 Object key = it.next();\r
 250                 System.out.print(" " + toHex2.transliterate((String)key) + ": ");\r
 251                 Set s = (Set) fromCaseFold.get(key);\r
 252                 Iterator it2 = s.iterator();\r
 253                 boolean first = true;\r
 254                 while (it2.hasNext()) {\r
 255                     if (first) {\r
 256                         first = false;\r
 257                     } else {\r
 258                         System.out.print(", ");\r
 259                     }\r
 260                     System.out.print(toHex2.transliterate((String)it2.next()));\r
 261                 }\r
 262                 System.out.println("");\r
 263             }\r
 264 \r
 265             System.out.println("\ntoCaseFold:");\r
 266             it = toCaseFold.keySet().iterator();\r
 267             while (it.hasNext()) {\r
 268                 String key = (String) it.next();\r
 269                 String value = (String) toCaseFold.get(key);\r
 270                 System.out.println(" " + toHex2.transliterate(key) + ": " + toHex2.transliterate(value));\r
 271             }            \r
 272         }\r
 273         \r
 274         // Now convert all those sets into linear arrays\r
 275         // We can't do this in place in Java, so make a temporary target array\r
 276         \r
 277         // Note: This could be transformed into a single array, with offsets into it.\r
 278         // Might be best choice in C.\r
 279         \r
 280         \r
 281         Map fromCaseFold2 = new HashMap();\r
 282         Iterator it = fromCaseFold.keySet().iterator();\r
 283         while (it.hasNext()) {\r
 284             Object key = it.next();\r
 285             Set s = (Set) fromCaseFold.get(key);\r
 286             String[] temp = new String[s.size()];\r
 287             s.toArray(temp);\r
 288             fromCaseFold2.put(key, temp);\r
 289         }\r
 290         fromCaseFold = fromCaseFold2;\r
 291 \r
 292         // We have processed everything, so the iterator will now work\r
 293         // The following is normally OFF. \r
 294         // It is here to generate (under the GENERATE flag) the static exception list.\r
 295         // It must be at the very end of initialization, so that the iterator is functional.\r
 296         // (easiest to do it that way)\r
 297             \r
 298         if (GENERATE) {\r
 299 \r
 300             // first get small set of items that have multiple characters\r
 301             \r
 302             Set multichars = new TreeSet();\r
 303             it = fromCaseFold.keySet().iterator();\r
 304             while (it.hasNext()) {\r
 305                 String key = (String) it.next();\r
 306                 if (UTF16.countCodePoint(key) < 2) continue;\r
 307                 multichars.add(key);\r
 308             }            \r
 309             \r
 310             // now we will go through each of them.\r
 311             \r
 312             CaseIterator ci = new CaseIterator();\r
 313             it = multichars.iterator();\r
 314             \r
 315             while (it.hasNext()) {\r
 316                 String key = (String) it.next();\r
 317                 \r
 318                 // here is a nasty complication. Take 'ffi' ligature. We\r
 319                 // can't just close it, since we would miss the combination\r
 320                 // that includes the 'fi' => "fi" ligature\r
 321                 // so first do a pass through, and add substring combinations\r
 322                 // we call this a 'partial closure'\r
 323                 \r
 324                 Set partialClosure = new TreeSet();\r
 325                 partialClosure.add(key);\r
 326                 \r
 327                 if (UTF16.countCodePoint(key) > 2) {\r
 328                     Iterator multiIt2 = multichars.iterator();\r
 329                     while (multiIt2.hasNext()) {\r
 330                         String otherKey = (String) multiIt2.next();\r
 331                         if (otherKey.length() >= key.length()) continue;\r
 332                         int pos = -1;\r
 333                         while (true) {\r
 334                             // The following is not completely general\r
 335                             // but works for the actual cased stuff,\r
 336                             // and should work for future characters, since we won't have\r
 337                             // more ligatures & other oddities.\r
 338                             pos = key.indexOf(otherKey, pos+1);\r
 339                             if (pos < 0) break;\r
 340                             int endPos = pos + otherKey.length();\r
 341                             // we know we have a proper substring,\r
 342                             // so get the combinations\r
 343                             String[] choices = (String[]) fromCaseFold.get(otherKey);\r
 344                             for (int ii = 0; ii < choices.length; ++ii) {\r
 345                                 String patchwork = key.substring(0, pos)\r
 346                                     + choices[ii]\r
 347                                     + key.substring(endPos);\r
 348                                 partialClosure.add(patchwork);\r
 349                             }\r
 350                         }\r
 351                     }\r
 352                 }\r
 353                 \r
 354                 // now, for each thing in the partial closure, get its\r
 355                 // case closure and add it to the final result.\r
 356                 \r
 357                 Set closure = new TreeSet(); // this will be the real closure\r
 358                 Iterator partialIt = partialClosure.iterator();\r
 359                 while (partialIt.hasNext()) {\r
 360                     String key2 = (String) partialIt.next();\r
 361                     ci.reset(key2);\r
 362                     for (String temp = ci.next(); temp != null; temp = ci.next()) {\r
 363                         closure.add(temp);\r
 364                     }\r
 365                     // form closure\r
 366                     /*String[] choices = (String[]) fromCaseFold.get(key2);\r
 367                     for (int i = 0; i < choices.length; ++i) {\r
 368                         ci.reset(choices[i]);\r
 369                         String temp;\r
 370                         while (null != (temp = ci.next())) {\r
 371                             closure.add(temp);\r
 372                         }\r
 373                     }\r
 374                     */\r
 375                 }\r
 376                 \r
 377                 // print it out, so that it can be cut and pasted back into this document.\r
 378                 \r
 379                 Iterator it2 = closure.iterator();\r
 380                 System.out.println("\t// " + toName.transliterate(key));\r
 381                 System.out.print("\t{\"" + toHex.transliterate(key) + "\",");\r
 382                 while (it2.hasNext()) {\r
 383                     String item = (String)it2.next();\r
 384                     System.out.print("\"" + toHex.transliterate(item) + "\",");\r
 385                 }\r
 386                 System.out.println("},");\r
 387             }\r
 388         }\r
 389     }\r
 390     \r
 391     // ============ PRIVATE CLASS DATA ============ \r
 392     \r
 393     // pieces that we will put together\r
 394     // is not changed during iteration\r
 395     private int count = 0;\r
 396     private String[][] variants;\r
 397     \r
 398     // state information, changes during iteration\r
 399     private boolean done = false;\r
 400     private int[] counts;\r
 401     \r
 402     // internal buffer for efficiency\r
 403     private StringBuffer nextBuffer = new StringBuffer();\r
 404     \r
 405     // ========================  \r
 406 \r
 407     /**\r
 408      * Reset to different source. Once reset, the iteration starts from the beginning.\r
 409      * @param source The string to get case variants for\r
 410      */\r
 411     public void reset(String source) {\r
 412         \r
 413         // allocate arrays to store pieces\r
 414         // using length might be slightly too long, but we don't care much\r
 415         \r
 416         counts = new int[source.length()];\r
 417         variants = new String[source.length()][];\r
 418         \r
 419         // walk through the source, and break up into pieces\r
 420         // each piece becomes an array of equivalent values\r
 421         // TODO: could optimized this later to coalesce all single string pieces\r
 422         \r
 423         String piece = null;\r
 424         count = 0;\r
 425         for (int i = 0; i < source.length(); i += piece.length()) {\r
 426             \r
 427             // find *longest* matching piece\r
 428             String caseFold = null;\r
 429             \r
 430             if (GENERATE) {\r
 431                 // do exactly one CP\r
 432                 piece = UTF16.valueOf(source, i);\r
 433                 caseFold = (String) toCaseFold.get(piece);\r
 434             } else {               \r
 435                 int max = i + maxLength;\r
 436                 if (max > source.length()) max = source.length();\r
 437                 for (int j = max; j > i; --j) {\r
 438                     piece = source.substring(i, j);\r
 439                     caseFold = (String) toCaseFold.get(piece);\r
 440                     if (caseFold != null) break;\r
 441                 }\r
 442             }\r
 443             \r
 444             // if we fail, pick one code point\r
 445             if (caseFold == null) {\r
 446                 piece = UTF16.valueOf(source, i);\r
 447                 variants[count++] = new String[] {piece}; // single item string\r
 448             } else {\r
 449                 variants[count++] = (String[])fromCaseFold.get(caseFold);\r
 450             }\r
 451         }\r
 452         reset();\r
 453     }\r
 454     \r
 455     /**\r
 456      * Restart the iteration from the beginning, but with same source\r
 457      */\r
 458     public void reset() {\r
 459         done = false;\r
 460         for (int i = 0; i < count; ++i) {\r
 461             counts[i] = 0;\r
 462         }\r
 463     }\r
 464     \r
 465     /**\r
 466      * Iterates through the case variants.\r
 467      * @return next case variant. Each variant will case-fold to the same value as the source will.\r
 468      * When the iteration is done, null is returned.\r
 469      */\r
 470     public String next() {\r
 471         \r
 472         if (done) return null;\r
 473         int i;\r
 474         \r
 475         // TODO Optimize so we keep the piece before and after the current position\r
 476         // so we don't have so much concatenation\r
 477         \r
 478         // get the result, a concatenation\r
 479         \r
 480         nextBuffer.setLength(0);\r
 481         for (i = 0; i < count; ++i) {\r
 482             nextBuffer.append(variants[i][counts[i]]);\r
 483         }\r
 484         \r
 485         // find the next right set of pieces to concatenate\r
 486         \r
 487         for (i = count-1; i >= 0; --i) {\r
 488             counts[i]++;\r
 489             if (counts[i] < variants[i].length) break;\r
 490             counts[i] = 0;\r
 491         }\r
 492         \r
 493         // if we go too far, bail\r
 494         \r
 495         if (i < 0) {\r
 496             done = true;\r
 497         }\r
 498         \r
 499         return nextBuffer.toString();            \r
 500     }\r
 501         \r
 502         \r
 503     /**\r
 504      * Temporary test, just to see how the stuff works.\r
 505      */\r
 506     static public void main(String[] args) {\r
 507         String[] testCases = {"fiss", "h\u03a3"};\r
 508         CaseIterator ci = new CaseIterator();\r
 509         \r
 510         for (int i = 0; i < testCases.length; ++i) {\r
 511             String item = testCases[i];\r
 512             System.out.println();\r
 513             System.out.println("Testing: " + toName.transliterate(item));\r
 514             System.out.println();\r
 515             ci.reset(item);\r
 516             int count = 0;\r
 517             for (String temp = ci.next(); temp != null; temp = ci.next()) {\r
 518                 System.out.println(toName.transliterate(temp));\r
 519                 count++;\r
 520             }\r
 521             System.out.println("Total: " + count);\r
 522         }\r
 523 \r
 524         // generate a list of all caseless characters -- characters whose\r
 525         // case closure is themselves.\r
 526 \r
 527         UnicodeSet caseless = new UnicodeSet();\r
 528 \r
 529         for (int i = 0; i <= 0x10FFFF; ++i) {\r
 530             String cp = UTF16.valueOf(i);\r
 531             ci.reset(cp);\r
 532             int count = 0;\r
 533             String fold = null;\r
 534             for (String temp = ci.next(); temp != null; temp = ci.next()) {\r
 535                 fold = temp;\r
 536                 if (++count > 1) break;\r
 537             }\r
 538             if (count==1 && fold.equals(cp)) {\r
 539                 caseless.add(i);\r
 540             }\r
 541         }\r
 542 \r
 543         System.out.println("caseless = " + caseless.toPattern(true));\r
 544 \r
 545         UnicodeSet not_lc = new UnicodeSet("[:^lc:]");\r
 546         \r
 547         UnicodeSet a = new UnicodeSet();\r
 548         a.set(not_lc);\r
 549         a.removeAll(caseless);\r
 550         System.out.println("[:^lc:] - caseless = " + a.toPattern(true));\r
 551 \r
 552         a.set(caseless);\r
 553         a.removeAll(not_lc);\r
 554         System.out.println("caseless - [:^lc:] = " + a.toPattern(true));\r
 555     }\r
 556 }\r