KickJava   Java API By Example, From Geeks To Geeks.

Java > Open Source Codes > org > apache > xerces > impl > xpath > regex > ParserForXMLSchema


1 /*
2  * Copyright 1999-2005 The Apache Software Foundation.
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 org.apache.xerces.impl.xpath.regex;
18
19 import java.util.Hashtable JavaDoc;
20 import java.util.Locale JavaDoc;
21
22 /**
23  * A regular expression parser for the XML Schema.
24  *
25  * @xerces.internal
26  *
27  * @author TAMURA Kent <kent@trl.ibm.co.jp>
28  * @version $Id: ParserForXMLSchema.java,v 1.9 2005/06/23 15:55:18 mrglavas Exp $
29  */
30 class ParserForXMLSchema extends RegexParser {
31
32     public ParserForXMLSchema() {
33         //this.setLocale(Locale.getDefault());
34 }
35     public ParserForXMLSchema(Locale JavaDoc locale) {
36         //this.setLocale(locale);
37 }
38
39     Token processCaret() throws ParseException {
40         this.next();
41         return Token.createChar('^');
42     }
43     Token processDollar() throws ParseException {
44         this.next();
45         return Token.createChar('$');
46      }
47     Token processLookahead() throws ParseException {
48         throw ex("parser.process.1", this.offset);
49     }
50     Token processNegativelookahead() throws ParseException {
51         throw ex("parser.process.1", this.offset);
52     }
53     Token processLookbehind() throws ParseException {
54         throw ex("parser.process.1", this.offset);
55     }
56     Token processNegativelookbehind() throws ParseException {
57         throw ex("parser.process.1", this.offset);
58     }
59     Token processBacksolidus_A() throws ParseException {
60         throw ex("parser.process.1", this.offset);
61     }
62     Token processBacksolidus_Z() throws ParseException {
63         throw ex("parser.process.1", this.offset);
64     }
65     Token processBacksolidus_z() throws ParseException {
66         throw ex("parser.process.1", this.offset);
67     }
68     Token processBacksolidus_b() throws ParseException {
69         throw ex("parser.process.1", this.offset);
70     }
71     Token processBacksolidus_B() throws ParseException {
72         throw ex("parser.process.1", this.offset);
73     }
74     Token processBacksolidus_lt() throws ParseException {
75         throw ex("parser.process.1", this.offset);
76     }
77     Token processBacksolidus_gt() throws ParseException {
78         throw ex("parser.process.1", this.offset);
79     }
80     Token processStar(Token tok) throws ParseException {
81         this.next();
82         return Token.createClosure(tok);
83     }
84     Token processPlus(Token tok) throws ParseException {
85         // X+ -> XX*
86 this.next();
87         return Token.createConcat(tok, Token.createClosure(tok));
88     }
89     Token processQuestion(Token tok) throws ParseException {
90         // X? -> X|
91 this.next();
92         Token par = Token.createUnion();
93         par.addChild(tok);
94         par.addChild(Token.createEmpty());
95         return par;
96     }
97     boolean checkQuestion(int off) {
98         return false;
99     }
100     Token processParen() throws ParseException {
101         this.next();
102         Token tok = Token.createParen(this.parseRegex(), 0);
103         if (this.read() != T_RPAREN) throw ex("parser.factor.1", this.offset-1);
104         this.next(); // Skips ')'
105 return tok;
106     }
107     Token processParen2() throws ParseException {
108         throw ex("parser.process.1", this.offset);
109     }
110     Token processCondition() throws ParseException {
111         throw ex("parser.process.1", this.offset);
112     }
113     Token processModifiers() throws ParseException {
114         throw ex("parser.process.1", this.offset);
115     }
116     Token processIndependent() throws ParseException {
117         throw ex("parser.process.1", this.offset);
118     }
119     Token processBacksolidus_c() throws ParseException {
120         this.next();
121         return this.getTokenForShorthand('c');
122     }
123     Token processBacksolidus_C() throws ParseException {
124         this.next();
125         return this.getTokenForShorthand('C');
126     }
127     Token processBacksolidus_i() throws ParseException {
128         this.next();
129         return this.getTokenForShorthand('i');
130     }
131     Token processBacksolidus_I() throws ParseException {
132         this.next();
133         return this.getTokenForShorthand('I');
134     }
135     Token processBacksolidus_g() throws ParseException {
136         throw this.ex("parser.process.1", this.offset-2);
137     }
138     Token processBacksolidus_X() throws ParseException {
139         throw ex("parser.process.1", this.offset-2);
140     }
141     Token processBackreference() throws ParseException {
142         throw ex("parser.process.1", this.offset-4);
143     }
144
145     int processCIinCharacterClass(RangeToken tok, int c) {
146         tok.mergeRanges(this.getTokenForShorthand(c));
147         return -1;
148     }
149
150
151     /**
152      * Parses a character-class-expression, not a character-class-escape.
153      *
154      * c-c-expression ::= '[' c-group ']'
155      * c-group ::= positive-c-group | negative-c-group | c-c-subtraction
156      * positive-c-group ::= (c-range | c-c-escape)+
157      * negative-c-group ::= '^' positive-c-group
158      * c-c-subtraction ::= (positive-c-group | negative-c-group) subtraction
159      * subtraction ::= '-' c-c-expression
160      * c-range ::= single-range | from-to-range
161      * single-range ::= multi-c-escape | category-c-escape | block-c-escape | <any XML char>
162      * cc-normal-c ::= <any character except [, ], \>
163      * from-to-range ::= cc-normal-c '-' cc-normal-c
164      *
165      * @param useNrage Ignored.
166      * @return This returns no NrageToken.
167      */
168     protected RangeToken parseCharacterClass(boolean useNrange) throws ParseException {
169         this.setContext(S_INBRACKETS);
170         this.next(); // '['
171 boolean nrange = false;
172         boolean wasDecoded = false; // used to detect if the last - was escaped.
173 RangeToken base = null;
174         RangeToken tok;
175         if (this.read() == T_CHAR && this.chardata == '^') {
176             nrange = true;
177             this.next(); // '^'
178 base = Token.createRange();
179             base.addRange(0, Token.UTF16_MAX);
180             tok = Token.createRange();
181         } else {
182             tok = Token.createRange();
183         }
184         int type;
185         boolean firstloop = true;
186         while ((type = this.read()) != T_EOF) { // Don't use 'cotinue' for this loop.
187
188             wasDecoded = false;
189             // single-range | from-to-range | subtraction
190 if (type == T_CHAR && this.chardata == ']' && !firstloop) {
191                 if (nrange) {
192                     base.subtractRanges(tok);
193                     tok = base;
194                 }
195                 break;
196             }
197             int c = this.chardata;
198             boolean end = false;
199             if (type == T_BACKSOLIDUS) {
200                 switch (c) {
201                   case 'd': case 'D':
202                   case 'w': case 'W':
203                   case 's': case 'S':
204                     tok.mergeRanges(this.getTokenForShorthand(c));
205                     end = true;
206                     break;
207
208                   case 'i': case 'I':
209                   case 'c': case 'C':
210                     c = this.processCIinCharacterClass(tok, c);
211                     if (c < 0) end = true;
212                     break;
213                     
214                   case 'p':
215                   case 'P':
216                     int pstart = this.offset;
217                     RangeToken tok2 = this.processBacksolidus_pP(c);
218                     if (tok2 == null) throw this.ex("parser.atom.5", pstart);
219                     tok.mergeRanges(tok2);
220                     end = true;
221                     break;
222                    
223                  case '-':
224                     c = this.decodeEscaped();
225                     wasDecoded = true;
226                     break;
227
228                   default:
229                     c = this.decodeEscaped();
230                 } // \ + c
231 } // backsolidus
232 else if (type == T_XMLSCHEMA_CC_SUBTRACTION && !firstloop) {
233                                                 // Subraction
234 if (nrange) {
235                     base.subtractRanges(tok);
236                     tok = base;
237                 }
238                 RangeToken range2 = this.parseCharacterClass(false);
239                 tok.subtractRanges(range2);
240                 if (this.read() != T_CHAR || this.chardata != ']')
241                     throw this.ex("parser.cc.5", this.offset);
242                 break; // Exit this loop
243 }
244             this.next();
245             if (!end) { // if not shorthands...
246 if (type == T_CHAR) {
247                     if (c == '[') throw this.ex("parser.cc.6", this.offset-2);
248                     if (c == ']') throw this.ex("parser.cc.7", this.offset-2);
249                     if (c == '-' && this.chardata == ']' && firstloop) throw this.ex("parser.cc.8", this.offset-2); // if regex = '[-]' then invalid
250 }
251                 if(c == '-' && this.chardata == '-' && this.read() != T_BACKSOLIDUS && !wasDecoded) {
252                     throw this.ex("parser.cc.8", this.offset-2);
253                 }
254                 if (this.read() != T_CHAR || this.chardata != '-') { // Here is no '-'.
255 tok.addRange(c, c);
256                 } else { // Found '-'
257 // Is this '-' is a from-to token??
258 this.next(); // Skips '-'
259 if ((type = this.read()) == T_EOF) throw this.ex("parser.cc.2", this.offset);
260                                                 // c '-' ']' -> '-' is a single-range.
261 if(type == T_CHAR && this.chardata == ']') { // if - is at the last position of the group
262 tok.addRange(c, c);
263                         tok.addRange('-', '-');
264                     }
265                     else if (type == T_XMLSCHEMA_CC_SUBTRACTION) {
266                         throw this.ex("parser.cc.8", this.offset-1);
267                     } else {
268                         
269                         int rangeend = this.chardata;
270                         if (type == T_CHAR) {
271                             if (rangeend == '[') throw this.ex("parser.cc.6", this.offset-1);
272                             if (rangeend == ']') throw this.ex("parser.cc.7", this.offset-1);
273                             if (rangeend == '-') throw this.ex("parser.cc.8", this.offset-2);
274                         }
275                         else if (type == T_BACKSOLIDUS)
276                             rangeend = this.decodeEscaped();
277                         this.next();
278
279                         if (c > rangeend) throw this.ex("parser.ope.3", this.offset-1);
280                         tok.addRange(c, rangeend);
281                     }
282                 }
283             }
284             firstloop = false;
285         }
286         if (this.read() == T_EOF)
287             throw this.ex("parser.cc.2", this.offset);
288         tok.sortRanges();
289         tok.compactRanges();
290         //tok.dumpRanges();
291 this.setContext(S_NORMAL);
292         this.next(); // Skips ']'
293
294         return tok;
295     }
296
297     protected RangeToken parseSetOperations() throws ParseException {
298         throw this.ex("parser.process.1", this.offset);
299     }
300  
301     Token getTokenForShorthand(int ch) {
302         switch (ch) {
303           case 'd':
304             return ParserForXMLSchema.getRange("xml:isDigit", true);
305           case 'D':
306             return ParserForXMLSchema.getRange("xml:isDigit", false);
307           case 'w':
308             return ParserForXMLSchema.getRange("xml:isWord", true);
309           case 'W':
310             return ParserForXMLSchema.getRange("xml:isWord", false);
311           case 's':
312             return ParserForXMLSchema.getRange("xml:isSpace", true);
313           case 'S':
314             return ParserForXMLSchema.getRange("xml:isSpace", false);
315           case 'c':
316             return ParserForXMLSchema.getRange("xml:isNameChar", true);
317           case 'C':
318             return ParserForXMLSchema.getRange("xml:isNameChar", false);
319           case 'i':
320             return ParserForXMLSchema.getRange("xml:isInitialNameChar", true);
321           case 'I':
322             return ParserForXMLSchema.getRange("xml:isInitialNameChar", false);
323           default:
324             throw new RuntimeException JavaDoc("Internal Error: shorthands: \\u"+Integer.toString(ch, 16));
325         }
326     }
327     int decodeEscaped() throws ParseException {
328         if (this.read() != T_BACKSOLIDUS) throw ex("parser.next.1", this.offset-1);
329         int c = this.chardata;
330         switch (c) {
331           case 'n': c = '\n'; break; // LINE FEED U+000A
332 case 'r': c = '\r'; break; // CRRIAGE RETURN U+000D
333 case 't': c = '\t'; break; // HORIZONTAL TABULATION U+0009
334 case '\\':
335           case '|':
336           case '.':
337           case '^':
338           case '-':
339           case '?':
340           case '*':
341           case '+':
342           case '{':
343           case '}':
344           case '(':
345           case ')':
346           case '[':
347           case ']':
348             break; // return actucal char
349 default:
350             throw ex("parser.process.1", this.offset-2);
351         }
352         return c;
353     }
354
355     static private Hashtable JavaDoc ranges = null;
356     static private Hashtable JavaDoc ranges2 = null;
357     static synchronized protected RangeToken getRange(String JavaDoc name, boolean positive) {
358         if (ranges == null) {
359             ranges = new Hashtable JavaDoc();
360             ranges2 = new Hashtable JavaDoc();
361
362             Token tok = Token.createRange();
363             setupRange(tok, SPACES);
364             ranges.put("xml:isSpace", tok);
365             ranges2.put("xml:isSpace", Token.complementRanges(tok));
366
367             tok = Token.createRange();
368             setupRange(tok, DIGITS);
369             ranges.put("xml:isDigit", tok);
370             ranges2.put("xml:isDigit", Token.complementRanges(tok));
371
372             tok = Token.createRange();
373             setupRange(tok, DIGITS);
374             ranges.put("xml:isDigit", tok);
375             ranges2.put("xml:isDigit", Token.complementRanges(tok));
376
377             tok = Token.createRange();
378             setupRange(tok, LETTERS);
379             tok.mergeRanges((Token)ranges.get("xml:isDigit"));
380             ranges.put("xml:isWord", tok);
381             ranges2.put("xml:isWord", Token.complementRanges(tok));
382
383             tok = Token.createRange();
384             setupRange(tok, NAMECHARS);
385             ranges.put("xml:isNameChar", tok);
386             ranges2.put("xml:isNameChar", Token.complementRanges(tok));
387
388             tok = Token.createRange();
389             setupRange(tok, LETTERS);
390             tok.addRange('_', '_');
391             tok.addRange(':', ':');
392             ranges.put("xml:isInitialNameChar", tok);
393             ranges2.put("xml:isInitialNameChar", Token.complementRanges(tok));
394         }
395         RangeToken tok = positive ? (RangeToken)ranges.get(name)
396             : (RangeToken)ranges2.get(name);
397         return tok;
398     }
399
400     static void setupRange(Token range, String JavaDoc src) {
401         int len = src.length();
402         for (int i = 0; i < len; i += 2)
403             range.addRange(src.charAt(i), src.charAt(i+1));
404     }
405
406     private static final String JavaDoc SPACES = "\t\n\r\r ";
407     private static final String JavaDoc NAMECHARS =
408         "\u002d\u002e\u0030\u003a\u0041\u005a\u005f\u005f\u0061\u007a\u00b7\u00b7\u00c0\u00d6"
409         +"\u00d8\u00f6\u00f8\u0131\u0134\u013e\u0141\u0148\u014a\u017e\u0180\u01c3\u01cd\u01f0"
410         +"\u01f4\u01f5\u01fa\u0217\u0250\u02a8\u02bb\u02c1\u02d0\u02d1\u0300\u0345\u0360\u0361"
411         +"\u0386\u038a\u038c\u038c\u038e\u03a1\u03a3\u03ce\u03d0\u03d6\u03da\u03da\u03dc\u03dc"
412         +"\u03de\u03de\u03e0\u03e0\u03e2\u03f3\u0401\u040c\u040e\u044f\u0451\u045c\u045e\u0481"
413         +"\u0483\u0486\u0490\u04c4\u04c7\u04c8\u04cb\u04cc\u04d0\u04eb\u04ee\u04f5\u04f8\u04f9"
414         +"\u0531\u0556\u0559\u0559\u0561\u0586\u0591\u05a1\u05a3\u05b9\u05bb\u05bd\u05bf\u05bf"
415         +"\u05c1\u05c2\u05c4\u05c4\u05d0\u05ea\u05f0\u05f2\u0621\u063a\u0640\u0652\u0660\u0669"
416         +"\u0670\u06b7\u06ba\u06be\u06c0\u06ce\u06d0\u06d3\u06d5\u06e8\u06ea\u06ed\u06f0\u06f9"
417         +"\u0901\u0903\u0905\u0939\u093c\u094d\u0951\u0954\u0958\u0963\u0966\u096f\u0981\u0983"
418         +"\u0985\u098c\u098f\u0990\u0993\u09a8\u09aa\u09b0\u09b2\u09b2\u09b6\u09b9\u09bc\u09bc"
419         +"\u09be\u09c4\u09c7\u09c8\u09cb\u09cd\u09d7\u09d7\u09dc\u09dd\u09df\u09e3\u09e6\u09f1"
420         +"\u0a02\u0a02\u0a05\u0a0a\u0a0f\u0a10\u0a13\u0a28\u0a2a\u0a30\u0a32\u0a33\u0a35\u0a36"
421         +"\u0a38\u0a39\u0a3c\u0a3c\u0a3e\u0a42\u0a47\u0a48\u0a4b\u0a4d\u0a59\u0a5c\u0a5e\u0a5e"
422         +"\u0a66\u0a74\u0a81\u0a83\u0a85\u0a8b\u0a8d\u0a8d\u0a8f\u0a91\u0a93\u0aa8\u0aaa\u0ab0"
423         +"\u0ab2\u0ab3\u0ab5\u0ab9\u0abc\u0ac5\u0ac7\u0ac9\u0acb\u0acd\u0ae0\u0ae0\u0ae6\u0aef"
424         +"\u0b01\u0b03\u0b05\u0b0c\u0b0f\u0b10\u0b13\u0b28\u0b2a\u0b30\u0b32\u0b33\u0b36\u0b39"
425         +"\u0b3c\u0b43\u0b47\u0b48\u0b4b\u0b4d\u0b56\u0b57\u0b5c\u0b5d\u0b5f\u0b61\u0b66\u0b6f"
426         +"\u0b82\u0b83\u0b85\u0b8a\u0b8e\u0b90\u0b92\u0b95\u0b99\u0b9a\u0b9c\u0b9c\u0b9e\u0b9f"
427         +"\u0ba3\u0ba4\u0ba8\u0baa\u0bae\u0bb5\u0bb7\u0bb9\u0bbe\u0bc2\u0bc6\u0bc8\u0bca\u0bcd"
428         +"\u0bd7\u0bd7\u0be7\u0bef\u0c01\u0c03\u0c05\u0c0c\u0c0e\u0c10\u0c12\u0c28\u0c2a\u0c33"
429         +"\u0c35\u0c39\u0c3e\u0c44\u0c46\u0c48\u0c4a\u0c4d\u0c55\u0c56\u0c60\u0c61\u0c66\u0c6f"
430         +"\u0c82\u0c83\u0c85\u0c8c\u0c8e\u0c90\u0c92\u0ca8\u0caa\u0cb3\u0cb5\u0cb9\u0cbe\u0cc4"
431         +"\u0cc6\u0cc8\u0cca\u0ccd\u0cd5\u0cd6\u0cde\u0cde\u0ce0\u0ce1\u0ce6\u0cef\u0d02\u0d03"
432         +"\u0d05\u0d0c\u0d0e\u0d10\u0d12\u0d28\u0d2a\u0d39\u0d3e\u0d43\u0d46\u0d48\u0d4a\u0d4d"
433         +"\u0d57\u0d57\u0d60\u0d61\u0d66\u0d6f\u0e01\u0e2e\u0e30\u0e3a\u0e40\u0e4e\u0e50\u0e59"
434         +"\u0e81\u0e82\u0e84\u0e84\u0e87\u0e88\u0e8a\u0e8a\u0e8d\u0e8d\u0e94\u0e97\u0e99\u0e9f"
435         +"\u0ea1\u0ea3\u0ea5\u0ea5\u0ea7\u0ea7\u0eaa\u0eab\u0ead\u0eae\u0eb0\u0eb9\u0ebb\u0ebd"
436         +"\u0ec0\u0ec4\u0ec6\u0ec6\u0ec8\u0ecd\u0ed0\u0ed9\u0f18\u0f19\u0f20\u0f29\u0f35\u0f35"
437         +"\u0f37\u0f37\u0f39\u0f39\u0f3e\u0f47\u0f49\u0f69\u0f71\u0f84\u0f86\u0f8b\u0f90\u0f95"
438         +"\u0f97\u0f97\u0f99\u0fad\u0fb1\u0fb7\u0fb9\u0fb9\u10a0\u10c5\u10d0\u10f6\u1100\u1100"
439         +"\u1102\u1103\u1105\u1107\u1109\u1109\u110b\u110c\u110e\u1112\u113c\u113c\u113e\u113e"
440         +"\u1140\u1140\u114c\u114c\u114e\u114e\u1150\u1150\u1154\u1155\u1159\u1159\u115f\u1161"
441         +"\u1163\u1163\u1165\u1165\u1167\u1167\u1169\u1169\u116d\u116e\u1172\u1173\u1175\u1175"
442         +"\u119e\u119e\u11a8\u11a8\u11ab\u11ab\u11ae\u11af\u11b7\u11b8\u11ba\u11ba\u11bc\u11c2"
443         +"\u11eb\u11eb\u11f0\u11f0\u11f9\u11f9\u1e00\u1e9b\u1ea0\u1ef9\u1f00\u1f15\u1f18\u1f1d"
444         +"\u1f20\u1f45\u1f48\u1f4d\u1f50\u1f57\u1f59\u1f59\u1f5b\u1f5b\u1f5d\u1f5d\u1f5f\u1f7d"
445         +"\u1f80\u1fb4\u1fb6\u1fbc\u1fbe\u1fbe\u1fc2\u1fc4\u1fc6\u1fcc\u1fd0\u1fd3\u1fd6\u1fdb"
446         +"\u1fe0\u1fec\u1ff2\u1ff4\u1ff6\u1ffc\u20d0\u20dc\u20e1\u20e1\u2126\u2126\u212a\u212b"
447         +"\u212e\u212e\u2180\u2182\u3005\u3005\u3007\u3007\u3021\u302f\u3031\u3035\u3041\u3094"
448         +"\u3099\u309a\u309d\u309e\u30a1\u30fa\u30fc\u30fe\u3105\u312c\u4e00\u9fa5\uac00\ud7a3"
449         +"";
450     private static final String JavaDoc LETTERS =
451         "\u0041\u005a\u0061\u007a\u00c0\u00d6\u00d8\u00f6\u00f8\u0131\u0134\u013e\u0141\u0148"
452         +"\u014a\u017e\u0180\u01c3\u01cd\u01f0\u01f4\u01f5\u01fa\u0217\u0250\u02a8\u02bb\u02c1"
453         +"\u0386\u0386\u0388\u038a\u038c\u038c\u038e\u03a1\u03a3\u03ce\u03d0\u03d6\u03da\u03da"
454         +"\u03dc\u03dc\u03de\u03de\u03e0\u03e0\u03e2\u03f3\u0401\u040c\u040e\u044f\u0451\u045c"
455         +"\u045e\u0481\u0490\u04c4\u04c7\u04c8\u04cb\u04cc\u04d0\u04eb\u04ee\u04f5\u04f8\u04f9"
456         +"\u0531\u0556\u0559\u0559\u0561\u0586\u05d0\u05ea\u05f0\u05f2\u0621\u063a\u0641\u064a"
457         +"\u0671\u06b7\u06ba\u06be\u06c0\u06ce\u06d0\u06d3\u06d5\u06d5\u06e5\u06e6\u0905\u0939"
458         +"\u093d\u093d\u0958\u0961\u0985\u098c\u098f\u0990\u0993\u09a8\u09aa\u09b0\u09b2\u09b2"
459         +"\u09b6\u09b9\u09dc\u09dd\u09df\u09e1\u09f0\u09f1\u0a05\u0a0a\u0a0f\u0a10\u0a13\u0a28"
460         +"\u0a2a\u0a30\u0a32\u0a33\u0a35\u0a36\u0a38\u0a39\u0a59\u0a5c\u0a5e\u0a5e\u0a72\u0a74"
461         +"\u0a85\u0a8b\u0a8d\u0a8d\u0a8f\u0a91\u0a93\u0aa8\u0aaa\u0ab0\u0ab2\u0ab3\u0ab5\u0ab9"
462         +"\u0abd\u0abd\u0ae0\u0ae0\u0b05\u0b0c\u0b0f\u0b10\u0b13\u0b28\u0b2a\u0b30\u0b32\u0b33"
463         +"\u0b36\u0b39\u0b3d\u0b3d\u0b5c\u0b5d\u0b5f\u0b61\u0b85\u0b8a\u0b8e\u0b90\u0b92\u0b95"
464         +"\u0b99\u0b9a\u0b9c\u0b9c\u0b9e\u0b9f\u0ba3\u0ba4\u0ba8\u0baa\u0bae\u0bb5\u0bb7\u0bb9"
465         +"\u0c05\u0c0c\u0c0e\u0c10\u0c12\u0c28\u0c2a\u0c33\u0c35\u0c39\u0c60\u0c61\u0c85\u0c8c"
466         +"\u0c8e\u0c90\u0c92\u0ca8\u0caa\u0cb3\u0cb5\u0cb9\u0cde\u0cde\u0ce0\u0ce1\u0d05\u0d0c"
467         +"\u0d0e\u0d10\u0d12\u0d28\u0d2a\u0d39\u0d60\u0d61\u0e01\u0e2e\u0e30\u0e30\u0e32\u0e33"
468         +"\u0e40\u0e45\u0e81\u0e82\u0e84\u0e84\u0e87\u0e88\u0e8a\u0e8a\u0e8d\u0e8d\u0e94\u0e97"
469         +"\u0e99\u0e9f\u0ea1\u0ea3\u0ea5\u0ea5\u0ea7\u0ea7\u0eaa\u0eab\u0ead\u0eae\u0eb0\u0eb0"
470         +"\u0eb2\u0eb3\u0ebd\u0ebd\u0ec0\u0ec4\u0f40\u0f47\u0f49\u0f69\u10a0\u10c5\u10d0\u10f6"
471         +"\u1100\u1100\u1102\u1103\u1105\u1107\u1109\u1109\u110b\u110c\u110e\u1112\u113c\u113c"
472         +"\u113e\u113e\u1140\u1140\u114c\u114c\u114e\u114e\u1150\u1150\u1154\u1155\u1159\u1159"
473         +"\u115f\u1161\u1163\u1163\u1165\u1165\u1167\u1167\u1169\u1169\u116d\u116e\u1172\u1173"
474         +"\u1175\u1175\u119e\u119e\u11a8\u11a8\u11ab\u11ab\u11ae\u11af\u11b7\u11b8\u11ba\u11ba"
475         +"\u11bc\u11c2\u11eb\u11eb\u11f0\u11f0\u11f9\u11f9\u1e00\u1e9b\u1ea0\u1ef9\u1f00\u1f15"
476         +"\u1f18\u1f1d\u1f20\u1f45\u1f48\u1f4d\u1f50\u1f57\u1f59\u1f59\u1f5b\u1f5b\u1f5d\u1f5d"
477         +"\u1f5f\u1f7d\u1f80\u1fb4\u1fb6\u1fbc\u1fbe\u1fbe\u1fc2\u1fc4\u1fc6\u1fcc\u1fd0\u1fd3"
478         +"\u1fd6\u1fdb\u1fe0\u1fec\u1ff2\u1ff4\u1ff6\u1ffc\u2126\u2126\u212a\u212b\u212e\u212e"
479         +"\u2180\u2182\u3007\u3007\u3021\u3029\u3041\u3094\u30a1\u30fa\u3105\u312c\u4e00\u9fa5"
480         +"\uac00\ud7a3";
481     private static final String JavaDoc DIGITS =
482         "\u0030\u0039\u0660\u0669\u06F0\u06F9\u0966\u096F\u09E6\u09EF\u0A66\u0A6F\u0AE6\u0AEF"
483         +"\u0B66\u0B6F\u0BE7\u0BEF\u0C66\u0C6F\u0CE6\u0CEF\u0D66\u0D6F\u0E50\u0E59\u0ED0\u0ED9"
484         +"\u0F20\u0F29";
485 }
486
Popular Tags