1 /*
2 * Copyright (c) 2003, 2018, Oracle and/or its affiliates. All rights reserved.
3 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
4 *
5 * This code is free software; you can redistribute it and/or modify it
6 * under the terms of the GNU General Public License version 2 only, as
7 * published by the Free Software Foundation. Oracle designates this
8 * particular file as subject to the "Classpath" exception as provided
9 * by Oracle in the LICENSE file that accompanied this code.
10 *
11 * This code is distributed in the hope that it will be useful, but WITHOUT
12 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
13 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
14 * version 2 for more details (a copy is included in the LICENSE file that
15 * accompanied this code).
16 *
17 * You should have received a copy of the GNU General Public License version
18 * 2 along with this work; if not, write to the Free Software Foundation,
19 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
20 *
21 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
22 * or visit www.oracle.com if you need additional information or have any
23 * questions.
24 */
25
26 package java.lang;
27
28 /**
29 * The CharacterData00 class encapsulates the large tables once found in
30 * java.lang.Character
31 */
32
33 class CharacterData00 extends CharacterData {
34 /* The character properties are currently encoded into 32 bits in the following manner:
35 1 bit mirrored property
36 4 bits directionality property
37 9 bits signed offset used for converting case
38 1 bit if 1, adding the signed offset converts the character to lowercase
39 1 bit if 1, subtracting the signed offset converts the character to uppercase
40 1 bit if 1, this character has a titlecase equivalent (possibly itself)
41 3 bits 0 may not be part of an identifier
42 1 ignorable control; may continue a Unicode identifier or Java identifier
43 2 may continue a Java identifier but not a Unicode identifier (unused)
44 3 may continue a Unicode identifier or Java identifier
45 4 is a Java whitespace character
46 5 may start or continue a Java identifier;
47 may continue but not start a Unicode identifier (underscores)
48 6 may start or continue a Java identifier but not a Unicode identifier ($)
49 7 may start or continue a Unicode identifier or Java identifier
50 Thus:
51 5, 6, 7 may start a Java identifier
52 1, 2, 3, 5, 6, 7 may continue a Java identifier
53 7 may start a Unicode identifier
54 1, 3, 5, 7 may continue a Unicode identifier
55 1 is ignorable within an identifier
56 4 is Java whitespace
57 2 bits 0 this character has no numeric property
58 1 adding the digit offset to the character code and then
59 masking with 0x1F will produce the desired numeric value
60 2 this character has a "strange" numeric value
61 3 a Java supradecimal digit: adding the digit offset to the
62 character code, then masking with 0x1F, then adding 10
63 will produce the desired numeric value
64 5 bits digit offset
65 5 bits character type
66
67 The encoding of character properties is subject to change at any time.
68 */
69
70 int getProperties(int ch) {
71 char offset = (char)ch;
72 int props = $$Lookup(offset);
73 return props;
74 }
75
76 int getPropertiesEx(int ch) {
77 char offset = (char)ch;
78 int props = $$LookupEx(offset);
79 return props;
80 }
81
82 int getType(int ch) {
83 int props = getProperties(ch);
84 return (props & $$maskType);
85 }
86
87 boolean isOtherLowercase(int ch) {
88 int props = getPropertiesEx(ch);
89 return (props & $$maskOtherLowercase) != 0;
90 }
91
92 boolean isOtherUppercase(int ch) {
93 int props = getPropertiesEx(ch);
94 return (props & $$maskOtherUppercase) != 0;
95 }
96
97 boolean isOtherAlphabetic(int ch) {
98 int props = getPropertiesEx(ch);
99 return (props & $$maskOtherAlphabetic) != 0;
100 }
101
102 boolean isIdeographic(int ch) {
103 int props = getPropertiesEx(ch);
104 return (props & $$maskIdeographic) != 0;
105 }
106
107 boolean isJavaIdentifierStart(int ch) {
108 int props = getProperties(ch);
109 return ((props & $$maskIdentifierInfo) >= $$lowJavaStart);
110 }
111
112 boolean isJavaIdentifierPart(int ch) {
113 int props = getProperties(ch);
114 return ((props & $$nonzeroJavaPart) != 0);
115 }
116
117 boolean isUnicodeIdentifierStart(int ch) {
118 int props = getProperties(ch);
119 return ((props & $$maskIdentifierInfo) == $$valueUnicodeStart);
120 }
121
122 boolean isUnicodeIdentifierPart(int ch) {
123 int props = getProperties(ch);
124 return ((props & $$maskUnicodePart) != 0);
125 }
126
127 boolean isIdentifierIgnorable(int ch) {
128 int props = getProperties(ch);
129 return ((props & $$maskIdentifierInfo) == $$valueIgnorable);
130 }
131
132 int toLowerCase(int ch) {
133 int mapChar = ch;
134 int val = getProperties(ch);
135
136 if ((val & $$maskLowerCase) != 0) {
137 if ((val & $$maskCaseOffset) == $$maskCaseOffset) {
138 switch(ch) {
139 case 0x0130: mapChar = 0x0069; break;
140 case 0x023A: mapChar = 0x2C65; break;
141 case 0x023E: mapChar = 0x2C66; break;
142 case 0x0412: mapChar = 0x1C80; break;
143 case 0x0414: mapChar = 0x1C81; break;
144 case 0x041E: mapChar = 0x1C82; break;
145 case 0x0421: mapChar = 0x1C83; break;
146 case 0x042A: mapChar = 0x1C86; break;
147 case 0x0462: mapChar = 0x1C87; break;
148 case 0x10A0: mapChar = 0x2D00; break;
149 case 0x10A1: mapChar = 0x2D01; break;
150 case 0x10A2: mapChar = 0x2D02; break;
151 case 0x10A3: mapChar = 0x2D03; break;
152 case 0x10A4: mapChar = 0x2D04; break;
153 case 0x10A5: mapChar = 0x2D05; break;
154 case 0x10A6: mapChar = 0x2D06; break;
155 case 0x10A7: mapChar = 0x2D07; break;
156 case 0x10A8: mapChar = 0x2D08; break;
157 case 0x10A9: mapChar = 0x2D09; break;
158 case 0x10AA: mapChar = 0x2D0A; break;
159 case 0x10AB: mapChar = 0x2D0B; break;
160 case 0x10AC: mapChar = 0x2D0C; break;
161 case 0x10AD: mapChar = 0x2D0D; break;
162 case 0x10AE: mapChar = 0x2D0E; break;
163 case 0x10AF: mapChar = 0x2D0F; break;
164 case 0x10B0: mapChar = 0x2D10; break;
165 case 0x10B1: mapChar = 0x2D11; break;
166 case 0x10B2: mapChar = 0x2D12; break;
167 case 0x10B3: mapChar = 0x2D13; break;
168 case 0x10B4: mapChar = 0x2D14; break;
169 case 0x10B5: mapChar = 0x2D15; break;
170 case 0x10B6: mapChar = 0x2D16; break;
171 case 0x10B7: mapChar = 0x2D17; break;
172 case 0x10B8: mapChar = 0x2D18; break;
173 case 0x10B9: mapChar = 0x2D19; break;
174 case 0x10BA: mapChar = 0x2D1A; break;
175 case 0x10BB: mapChar = 0x2D1B; break;
176 case 0x10BC: mapChar = 0x2D1C; break;
177 case 0x10BD: mapChar = 0x2D1D; break;
178 case 0x10BE: mapChar = 0x2D1E; break;
179 case 0x10BF: mapChar = 0x2D1F; break;
180 case 0x10C0: mapChar = 0x2D20; break;
181 case 0x10C1: mapChar = 0x2D21; break;
182 case 0x10C2: mapChar = 0x2D22; break;
183 case 0x10C3: mapChar = 0x2D23; break;
184 case 0x10C4: mapChar = 0x2D24; break;
185 case 0x10C5: mapChar = 0x2D25; break;
186 case 0x10C7: mapChar = 0x2D27; break;
187 case 0x10CD: mapChar = 0x2D2D; break;
188 case 0x13A0: mapChar = 0xAB70; break;
189 case 0x13A1: mapChar = 0xAB71; break;
190 case 0x13A2: mapChar = 0xAB72; break;
191 case 0x13A3: mapChar = 0xAB73; break;
192 case 0x13A4: mapChar = 0xAB74; break;
193 case 0x13A5: mapChar = 0xAB75; break;
194 case 0x13A6: mapChar = 0xAB76; break;
195 case 0x13A7: mapChar = 0xAB77; break;
196 case 0x13A8: mapChar = 0xAB78; break;
197 case 0x13A9: mapChar = 0xAB79; break;
198 case 0x13AA: mapChar = 0xAB7A; break;
199 case 0x13AB: mapChar = 0xAB7B; break;
200 case 0x13AC: mapChar = 0xAB7C; break;
201 case 0x13AD: mapChar = 0xAB7D; break;
202 case 0x13AE: mapChar = 0xAB7E; break;
203 case 0x13AF: mapChar = 0xAB7F; break;
204 case 0x13B0: mapChar = 0xAB80; break;
205 case 0x13B1: mapChar = 0xAB81; break;
206 case 0x13B2: mapChar = 0xAB82; break;
207 case 0x13B3: mapChar = 0xAB83; break;
208 case 0x13B4: mapChar = 0xAB84; break;
209 case 0x13B5: mapChar = 0xAB85; break;
210 case 0x13B6: mapChar = 0xAB86; break;
211 case 0x13B7: mapChar = 0xAB87; break;
212 case 0x13B8: mapChar = 0xAB88; break;
213 case 0x13B9: mapChar = 0xAB89; break;
214 case 0x13BA: mapChar = 0xAB8A; break;
215 case 0x13BB: mapChar = 0xAB8B; break;
216 case 0x13BC: mapChar = 0xAB8C; break;
217 case 0x13BD: mapChar = 0xAB8D; break;
218 case 0x13BE: mapChar = 0xAB8E; break;
219 case 0x13BF: mapChar = 0xAB8F; break;
220 case 0x13C0: mapChar = 0xAB90; break;
221 case 0x13C1: mapChar = 0xAB91; break;
222 case 0x13C2: mapChar = 0xAB92; break;
223 case 0x13C3: mapChar = 0xAB93; break;
224 case 0x13C4: mapChar = 0xAB94; break;
225 case 0x13C5: mapChar = 0xAB95; break;
226 case 0x13C6: mapChar = 0xAB96; break;
227 case 0x13C7: mapChar = 0xAB97; break;
228 case 0x13C8: mapChar = 0xAB98; break;
229 case 0x13C9: mapChar = 0xAB99; break;
230 case 0x13CA: mapChar = 0xAB9A; break;
231 case 0x13CB: mapChar = 0xAB9B; break;
232 case 0x13CC: mapChar = 0xAB9C; break;
233 case 0x13CD: mapChar = 0xAB9D; break;
234 case 0x13CE: mapChar = 0xAB9E; break;
235 case 0x13CF: mapChar = 0xAB9F; break;
236 case 0x13D0: mapChar = 0xABA0; break;
237 case 0x13D1: mapChar = 0xABA1; break;
238 case 0x13D2: mapChar = 0xABA2; break;
239 case 0x13D3: mapChar = 0xABA3; break;
240 case 0x13D4: mapChar = 0xABA4; break;
241 case 0x13D5: mapChar = 0xABA5; break;
242 case 0x13D6: mapChar = 0xABA6; break;
243 case 0x13D7: mapChar = 0xABA7; break;
244 case 0x13D8: mapChar = 0xABA8; break;
245 case 0x13D9: mapChar = 0xABA9; break;
246 case 0x13DA: mapChar = 0xABAA; break;
247 case 0x13DB: mapChar = 0xABAB; break;
248 case 0x13DC: mapChar = 0xABAC; break;
249 case 0x13DD: mapChar = 0xABAD; break;
250 case 0x13DE: mapChar = 0xABAE; break;
251 case 0x13DF: mapChar = 0xABAF; break;
252 case 0x13E0: mapChar = 0xABB0; break;
253 case 0x13E1: mapChar = 0xABB1; break;
254 case 0x13E2: mapChar = 0xABB2; break;
255 case 0x13E3: mapChar = 0xABB3; break;
256 case 0x13E4: mapChar = 0xABB4; break;
257 case 0x13E5: mapChar = 0xABB5; break;
258 case 0x13E6: mapChar = 0xABB6; break;
259 case 0x13E7: mapChar = 0xABB7; break;
260 case 0x13E8: mapChar = 0xABB8; break;
261 case 0x13E9: mapChar = 0xABB9; break;
262 case 0x13EA: mapChar = 0xABBA; break;
263 case 0x13EB: mapChar = 0xABBB; break;
264 case 0x13EC: mapChar = 0xABBC; break;
265 case 0x13ED: mapChar = 0xABBD; break;
266 case 0x13EE: mapChar = 0xABBE; break;
267 case 0x13EF: mapChar = 0xABBF; break;
268 case 0x1E9E: mapChar = 0x00DF; break;
269 case 0x1F88: mapChar = 0x1F80; break;
270 case 0x1F89: mapChar = 0x1F81; break;
271 case 0x1F8A: mapChar = 0x1F82; break;
272 case 0x1F8B: mapChar = 0x1F83; break;
273 case 0x1F8C: mapChar = 0x1F84; break;
274 case 0x1F8D: mapChar = 0x1F85; break;
275 case 0x1F8E: mapChar = 0x1F86; break;
276 case 0x1F8F: mapChar = 0x1F87; break;
277 case 0x1F98: mapChar = 0x1F90; break;
278 case 0x1F99: mapChar = 0x1F91; break;
279 case 0x1F9A: mapChar = 0x1F92; break;
280 case 0x1F9B: mapChar = 0x1F93; break;
281 case 0x1F9C: mapChar = 0x1F94; break;
282 case 0x1F9D: mapChar = 0x1F95; break;
283 case 0x1F9E: mapChar = 0x1F96; break;
284 case 0x1F9F: mapChar = 0x1F97; break;
285 case 0x1FA8: mapChar = 0x1FA0; break;
286 case 0x1FA9: mapChar = 0x1FA1; break;
287 case 0x1FAA: mapChar = 0x1FA2; break;
288 case 0x1FAB: mapChar = 0x1FA3; break;
289 case 0x1FAC: mapChar = 0x1FA4; break;
290 case 0x1FAD: mapChar = 0x1FA5; break;
291 case 0x1FAE: mapChar = 0x1FA6; break;
292 case 0x1FAF: mapChar = 0x1FA7; break;
293 case 0x1FBC: mapChar = 0x1FB3; break;
294 case 0x1FCC: mapChar = 0x1FC3; break;
295 case 0x1FFC: mapChar = 0x1FF3; break;
296 case 0x2126: mapChar = 0x03C9; break;
297 case 0x212A: mapChar = 0x006B; break;
298 case 0x212B: mapChar = 0x00E5; break;
299 case 0x2C62: mapChar = 0x026B; break;
300 case 0x2C63: mapChar = 0x1D7D; break;
301 case 0x2C64: mapChar = 0x027D; break;
302 case 0x2C6D: mapChar = 0x0251; break;
303 case 0x2C6E: mapChar = 0x0271; break;
304 case 0x2C6F: mapChar = 0x0250; break;
305 case 0x2C70: mapChar = 0x0252; break;
306 case 0x2C7E: mapChar = 0x023F; break;
307 case 0x2C7F: mapChar = 0x0240; break;
308 case 0xA64A: mapChar = 0x1C88; break;
309 case 0xA77D: mapChar = 0x1D79; break;
310 case 0xA78D: mapChar = 0x0265; break;
311 case 0xA7AA: mapChar = 0x0266; break;
312 case 0xA7AB: mapChar = 0x025C; break;
313 case 0xA7AC: mapChar = 0x0261; break;
314 case 0xA7AD: mapChar = 0x026C; break;
315 case 0xA7AE: mapChar = 0x026A; break;
316 case 0xA7B0: mapChar = 0x029E; break;
317 case 0xA7B1: mapChar = 0x0287; break;
318 case 0xA7B2: mapChar = 0x029D; break;
319 case 0xA7B3: mapChar = 0xAB53; break;
320 // default mapChar is already set, so no
321 // need to redo it here.
322 // default : mapChar = ch;
323 }
324 }
325 else {
326 int offset = val << $$shiftCaseOffsetSign >> ($$shiftCaseOffsetSign+$$shiftCaseOffset);
327 mapChar = ch + offset;
328 }
329 }
330 return mapChar;
331 }
332
333 int toUpperCase(int ch) {
334 int mapChar = ch;
335 int val = getProperties(ch);
336
337 if ((val & $$maskUpperCase) != 0) {
338 if ((val & $$maskCaseOffset) == $$maskCaseOffset) {
339 switch(ch) {
340 case 0x017F: mapChar = 0x0053; break;
341 case 0x023F: mapChar = 0x2C7E; break;
342 case 0x0240: mapChar = 0x2C7F; break;
343 case 0x0250: mapChar = 0x2C6F; break;
344 case 0x0251: mapChar = 0x2C6D; break;
345 case 0x0252: mapChar = 0x2C70; break;
346 case 0x025C: mapChar = 0xA7AB; break;
347 case 0x0261: mapChar = 0xA7AC; break;
348 case 0x0265: mapChar = 0xA78D; break;
349 case 0x0266: mapChar = 0xA7AA; break;
350 case 0x026A: mapChar = 0xA7AE; break;
351 case 0x026B: mapChar = 0x2C62; break;
352 case 0x026C: mapChar = 0xA7AD; break;
353 case 0x0271: mapChar = 0x2C6E; break;
354 case 0x027D: mapChar = 0x2C64; break;
355 case 0x0287: mapChar = 0xA7B1; break;
356 case 0x029D: mapChar = 0xA7B2; break;
357 case 0x029E: mapChar = 0xA7B0; break;
358 case 0x1C80: mapChar = 0x0412; break;
359 case 0x1C81: mapChar = 0x0414; break;
360 case 0x1C82: mapChar = 0x041E; break;
361 case 0x1C83: mapChar = 0x0421; break;
362 case 0x1C84: mapChar = 0x0422; break;
363 case 0x1C85: mapChar = 0x0422; break;
364 case 0x1C86: mapChar = 0x042A; break;
365 case 0x1C87: mapChar = 0x0462; break;
366 case 0x1C88: mapChar = 0xA64A; break;
367 case 0x1D79: mapChar = 0xA77D; break;
368 case 0x1D7D: mapChar = 0x2C63; break;
369 case 0x1F80: mapChar = 0x1F88; break;
370 case 0x1F81: mapChar = 0x1F89; break;
371 case 0x1F82: mapChar = 0x1F8A; break;
372 case 0x1F83: mapChar = 0x1F8B; break;
373 case 0x1F84: mapChar = 0x1F8C; break;
374 case 0x1F85: mapChar = 0x1F8D; break;
375 case 0x1F86: mapChar = 0x1F8E; break;
376 case 0x1F87: mapChar = 0x1F8F; break;
377 case 0x1F90: mapChar = 0x1F98; break;
378 case 0x1F91: mapChar = 0x1F99; break;
379 case 0x1F92: mapChar = 0x1F9A; break;
380 case 0x1F93: mapChar = 0x1F9B; break;
381 case 0x1F94: mapChar = 0x1F9C; break;
382 case 0x1F95: mapChar = 0x1F9D; break;
383 case 0x1F96: mapChar = 0x1F9E; break;
384 case 0x1F97: mapChar = 0x1F9F; break;
385 case 0x1FA0: mapChar = 0x1FA8; break;
386 case 0x1FA1: mapChar = 0x1FA9; break;
387 case 0x1FA2: mapChar = 0x1FAA; break;
388 case 0x1FA3: mapChar = 0x1FAB; break;
389 case 0x1FA4: mapChar = 0x1FAC; break;
390 case 0x1FA5: mapChar = 0x1FAD; break;
391 case 0x1FA6: mapChar = 0x1FAE; break;
392 case 0x1FA7: mapChar = 0x1FAF; break;
393 case 0x1FB3: mapChar = 0x1FBC; break;
394 case 0x1FBE: mapChar = 0x0399; break;
395 case 0x1FC3: mapChar = 0x1FCC; break;
396 case 0x1FF3: mapChar = 0x1FFC; break;
397 case 0x2C65: mapChar = 0x023A; break;
398 case 0x2C66: mapChar = 0x023E; break;
399 case 0x2D00: mapChar = 0x10A0; break;
400 case 0x2D01: mapChar = 0x10A1; break;
401 case 0x2D02: mapChar = 0x10A2; break;
402 case 0x2D03: mapChar = 0x10A3; break;
403 case 0x2D04: mapChar = 0x10A4; break;
404 case 0x2D05: mapChar = 0x10A5; break;
405 case 0x2D06: mapChar = 0x10A6; break;
406 case 0x2D07: mapChar = 0x10A7; break;
407 case 0x2D08: mapChar = 0x10A8; break;
408 case 0x2D09: mapChar = 0x10A9; break;
409 case 0x2D0A: mapChar = 0x10AA; break;
410 case 0x2D0B: mapChar = 0x10AB; break;
411 case 0x2D0C: mapChar = 0x10AC; break;
412 case 0x2D0D: mapChar = 0x10AD; break;
413 case 0x2D0E: mapChar = 0x10AE; break;
414 case 0x2D0F: mapChar = 0x10AF; break;
415 case 0x2D10: mapChar = 0x10B0; break;
416 case 0x2D11: mapChar = 0x10B1; break;
417 case 0x2D12: mapChar = 0x10B2; break;
418 case 0x2D13: mapChar = 0x10B3; break;
419 case 0x2D14: mapChar = 0x10B4; break;
420 case 0x2D15: mapChar = 0x10B5; break;
421 case 0x2D16: mapChar = 0x10B6; break;
422 case 0x2D17: mapChar = 0x10B7; break;
423 case 0x2D18: mapChar = 0x10B8; break;
424 case 0x2D19: mapChar = 0x10B9; break;
425 case 0x2D1A: mapChar = 0x10BA; break;
426 case 0x2D1B: mapChar = 0x10BB; break;
427 case 0x2D1C: mapChar = 0x10BC; break;
428 case 0x2D1D: mapChar = 0x10BD; break;
429 case 0x2D1E: mapChar = 0x10BE; break;
430 case 0x2D1F: mapChar = 0x10BF; break;
431 case 0x2D20: mapChar = 0x10C0; break;
432 case 0x2D21: mapChar = 0x10C1; break;
433 case 0x2D22: mapChar = 0x10C2; break;
434 case 0x2D23: mapChar = 0x10C3; break;
435 case 0x2D24: mapChar = 0x10C4; break;
436 case 0x2D25: mapChar = 0x10C5; break;
437 case 0x2D27: mapChar = 0x10C7; break;
438 case 0x2D2D: mapChar = 0x10CD; break;
439 case 0xAB53: mapChar = 0xA7B3; break;
440 case 0xAB70: mapChar = 0x13A0; break;
441 case 0xAB71: mapChar = 0x13A1; break;
442 case 0xAB72: mapChar = 0x13A2; break;
443 case 0xAB73: mapChar = 0x13A3; break;
444 case 0xAB74: mapChar = 0x13A4; break;
445 case 0xAB75: mapChar = 0x13A5; break;
446 case 0xAB76: mapChar = 0x13A6; break;
447 case 0xAB77: mapChar = 0x13A7; break;
448 case 0xAB78: mapChar = 0x13A8; break;
449 case 0xAB79: mapChar = 0x13A9; break;
450 case 0xAB7A: mapChar = 0x13AA; break;
451 case 0xAB7B: mapChar = 0x13AB; break;
452 case 0xAB7C: mapChar = 0x13AC; break;
453 case 0xAB7D: mapChar = 0x13AD; break;
454 case 0xAB7E: mapChar = 0x13AE; break;
455 case 0xAB7F: mapChar = 0x13AF; break;
456 case 0xAB80: mapChar = 0x13B0; break;
457 case 0xAB81: mapChar = 0x13B1; break;
458 case 0xAB82: mapChar = 0x13B2; break;
459 case 0xAB83: mapChar = 0x13B3; break;
460 case 0xAB84: mapChar = 0x13B4; break;
461 case 0xAB85: mapChar = 0x13B5; break;
462 case 0xAB86: mapChar = 0x13B6; break;
463 case 0xAB87: mapChar = 0x13B7; break;
464 case 0xAB88: mapChar = 0x13B8; break;
465 case 0xAB89: mapChar = 0x13B9; break;
466 case 0xAB8A: mapChar = 0x13BA; break;
467 case 0xAB8B: mapChar = 0x13BB; break;
468 case 0xAB8C: mapChar = 0x13BC; break;
469 case 0xAB8D: mapChar = 0x13BD; break;
470 case 0xAB8E: mapChar = 0x13BE; break;
471 case 0xAB8F: mapChar = 0x13BF; break;
472 case 0xAB90: mapChar = 0x13C0; break;
473 case 0xAB91: mapChar = 0x13C1; break;
474 case 0xAB92: mapChar = 0x13C2; break;
475 case 0xAB93: mapChar = 0x13C3; break;
476 case 0xAB94: mapChar = 0x13C4; break;
477 case 0xAB95: mapChar = 0x13C5; break;
478 case 0xAB96: mapChar = 0x13C6; break;
479 case 0xAB97: mapChar = 0x13C7; break;
480 case 0xAB98: mapChar = 0x13C8; break;
481 case 0xAB99: mapChar = 0x13C9; break;
482 case 0xAB9A: mapChar = 0x13CA; break;
483 case 0xAB9B: mapChar = 0x13CB; break;
484 case 0xAB9C: mapChar = 0x13CC; break;
485 case 0xAB9D: mapChar = 0x13CD; break;
486 case 0xAB9E: mapChar = 0x13CE; break;
487 case 0xAB9F: mapChar = 0x13CF; break;
488 case 0xABA0: mapChar = 0x13D0; break;
489 case 0xABA1: mapChar = 0x13D1; break;
490 case 0xABA2: mapChar = 0x13D2; break;
491 case 0xABA3: mapChar = 0x13D3; break;
492 case 0xABA4: mapChar = 0x13D4; break;
493 case 0xABA5: mapChar = 0x13D5; break;
494 case 0xABA6: mapChar = 0x13D6; break;
495 case 0xABA7: mapChar = 0x13D7; break;
496 case 0xABA8: mapChar = 0x13D8; break;
497 case 0xABA9: mapChar = 0x13D9; break;
498 case 0xABAA: mapChar = 0x13DA; break;
499 case 0xABAB: mapChar = 0x13DB; break;
500 case 0xABAC: mapChar = 0x13DC; break;
501 case 0xABAD: mapChar = 0x13DD; break;
502 case 0xABAE: mapChar = 0x13DE; break;
503 case 0xABAF: mapChar = 0x13DF; break;
504 case 0xABB0: mapChar = 0x13E0; break;
505 case 0xABB1: mapChar = 0x13E1; break;
506 case 0xABB2: mapChar = 0x13E2; break;
507 case 0xABB3: mapChar = 0x13E3; break;
508 case 0xABB4: mapChar = 0x13E4; break;
509 case 0xABB5: mapChar = 0x13E5; break;
510 case 0xABB6: mapChar = 0x13E6; break;
511 case 0xABB7: mapChar = 0x13E7; break;
512 case 0xABB8: mapChar = 0x13E8; break;
513 case 0xABB9: mapChar = 0x13E9; break;
514 case 0xABBA: mapChar = 0x13EA; break;
515 case 0xABBB: mapChar = 0x13EB; break;
516 case 0xABBC: mapChar = 0x13EC; break;
517 case 0xABBD: mapChar = 0x13ED; break;
518 case 0xABBE: mapChar = 0x13EE; break;
519 case 0xABBF: mapChar = 0x13EF; break;
520 // ch must have a 1:M case mapping, but we
521 // can't handle it here. Return ch.
522 // since mapChar is already set, no need
523 // to redo it here.
524 //default : mapChar = ch;
525 }
526 }
527 else {
528 int offset = val << $$shiftCaseOffsetSign >> ($$shiftCaseOffsetSign+$$shiftCaseOffset);
529 mapChar = ch - offset;
530 }
531 }
532 return mapChar;
533 }
534
535 int toTitleCase(int ch) {
536 int mapChar = ch;
537 int val = getProperties(ch);
538
539 if ((val & $$maskTitleCase) != 0) {
540 // There is a titlecase equivalent. Perform further checks:
541 if ((val & $$maskUpperCase) == 0) {
542 // The character does not have an uppercase equivalent, so it must
543 // already be uppercase; so add 1 to get the titlecase form.
544 mapChar = ch + 1;
545 }
546 else if ((val & $$maskLowerCase) == 0) {
547 // The character does not have a lowercase equivalent, so it must
548 // already be lowercase; so subtract 1 to get the titlecase form.
549 mapChar = ch - 1;
550 }
551 // else {
552 // The character has both an uppercase equivalent and a lowercase
553 // equivalent, so it must itself be a titlecase form; return it.
554 // return ch;
555 //}
556 }
557 else if ((val & $$maskUpperCase) != 0) {
558 // This character has no titlecase equivalent but it does have an
559 // uppercase equivalent, so use that (subtract the signed case offset).
560 mapChar = toUpperCase(ch);
561 }
562 return mapChar;
563 }
564
565 int digit(int ch, int radix) {
566 int value = -1;
567 if (radix >= Character.MIN_RADIX && radix <= Character.MAX_RADIX) {
568 int val = getProperties(ch);
569 int kind = val & $$maskType;
570 if (kind == Character.DECIMAL_DIGIT_NUMBER) {
571 value = ch + ((val & $$maskDigitOffset) >> $$shiftDigitOffset) & $$maskDigit;
572 }
573 else if ((val & $$maskNumericType) == $$valueJavaSupradecimal) {
574 // Java supradecimal digit
575 value = (ch + ((val & $$maskDigitOffset) >> $$shiftDigitOffset) & $$maskDigit) + 10;
576 }
577 }
578 return (value < radix) ? value : -1;
579 }
580
581 int getNumericValue(int ch) {
582 int val = getProperties(ch);
583 int retval = -1;
584
585 switch (val & $$maskNumericType) {
586 default: // cannot occur
587 case ($$valueNotNumeric): // not numeric
588 retval = -1;
589 break;
590 case ($$valueDigit): // simple numeric
591 retval = ch + ((val & $$maskDigitOffset) >> $$shiftDigitOffset) & $$maskDigit;
592 break;
593 case ($$valueStrangeNumeric) : // "strange" numeric
594 switch (ch) {
595 case 0x0BF1: retval = 100; break; // TAMIL NUMBER ONE HUNDRED
596 case 0x0BF2: retval = 1000; break; // TAMIL NUMBER ONE THOUSAND
597 case 0x0D71: retval = 100; break; // MALAYALAM NUMBER ONE HUNDRED
598 case 0x0D72: retval = 1000; break; // MALAYALAM NUMBER ONE THOUSAND
599 case 0x1375: retval = 40; break; // ETHIOPIC NUMBER FORTY
600 case 0x1376: retval = 50; break; // ETHIOPIC NUMBER FIFTY
601 case 0x1377: retval = 60; break; // ETHIOPIC NUMBER SIXTY
602 case 0x1378: retval = 70; break; // ETHIOPIC NUMBER SEVENTY
603 case 0x1379: retval = 80; break; // ETHIOPIC NUMBER EIGHTY
604 case 0x137A: retval = 90; break; // ETHIOPIC NUMBER NINETY
605 case 0x137B: retval = 100; break; // ETHIOPIC NUMBER HUNDRED
606 case 0x137C: retval = 10000; break; // ETHIOPIC NUMBER TEN THOUSAND
607 case 0x215F: retval = 1; break; // FRACTION NUMERATOR ONE
608 case 0x216C: retval = 50; break; // ROMAN NUMERAL FIFTY
609 case 0x216D: retval = 100; break; // ROMAN NUMERAL ONE HUNDRED
610 case 0x216E: retval = 500; break; // ROMAN NUMERAL FIVE HUNDRED
611 case 0x216F: retval = 1000; break; // ROMAN NUMERAL ONE THOUSAND
612 case 0x217C: retval = 50; break; // SMALL ROMAN NUMERAL FIFTY
613 case 0x217D: retval = 100; break; // SMALL ROMAN NUMERAL ONE HUNDRED
614 case 0x217E: retval = 500; break; // SMALL ROMAN NUMERAL FIVE HUNDRED
615 case 0x217F: retval = 1000; break; // SMALL ROMAN NUMERAL ONE THOUSAND
616 case 0x2180: retval = 1000; break; // ROMAN NUMERAL ONE THOUSAND C D
617 case 0x2181: retval = 5000; break; // ROMAN NUMERAL FIVE THOUSAND
618 case 0x2182: retval = 10000; break; // ROMAN NUMERAL TEN THOUSAND
619 case 0x2186: retval = 50; break; // ROMAN NUMERAL FIFTY EARLY FORM
620 case 0x2187: retval = 50000; break; // ROMAN NUMERAL FIFTY THOUSAND
621 case 0x2188: retval = 100000; break; // ROMAN NUMERAL ONE HUNDRED THOUSAND
622 case 0x324B: retval = 40; break; // CIRCLED NUMBER FORTY ON BLACK SQUARE
623 case 0x324C: retval = 50; break; // CIRCLED NUMBER FIFTY ON BLACK SQUARE
624 case 0x324D: retval = 60; break; // CIRCLED NUMBER SIXTY ON BLACK SQUARE
625 case 0x324E: retval = 70; break; // CIRCLED NUMBER SEVENTY ON BLACK SQUARE
626 case 0x324F: retval = 80; break; // CIRCLED NUMBER EIGHTY ON BLACK SQUARE
627 case 0x325C: retval = 32; break; // CIRCLED NUMBER THIRTY TWO
628 case 0x325D: retval = 33; break; // CIRCLED NUMBER THIRTY THREE
629 case 0x325E: retval = 34; break; // CIRCLED NUMBER THIRTY FOUR
630 case 0x325F: retval = 35; break; // CIRCLED NUMBER THIRTY FIVE
631 case 0x32B1: retval = 36; break; // CIRCLED NUMBER THIRTY SIX
632 case 0x32B2: retval = 37; break; // CIRCLED NUMBER THIRTY SEVEN
633 case 0x32B3: retval = 38; break; // CIRCLED NUMBER THIRTY EIGHT
634 case 0x32B4: retval = 39; break; // CIRCLED NUMBER THIRTY NINE
635 case 0x32B5: retval = 40; break; // CIRCLED NUMBER FORTY
636 case 0x32B6: retval = 41; break; // CIRCLED NUMBER FORTY ONE
637 case 0x32B7: retval = 42; break; // CIRCLED NUMBER FORTY TWO
638 case 0x32B8: retval = 43; break; // CIRCLED NUMBER FORTY THREE
639 case 0x32B9: retval = 44; break; // CIRCLED NUMBER FORTY FOUR
640 case 0x32BA: retval = 45; break; // CIRCLED NUMBER FORTY FIVE
641 case 0x32BB: retval = 46; break; // CIRCLED NUMBER FORTY SIX
642 case 0x32BC: retval = 47; break; // CIRCLED NUMBER FORTY SEVEN
643 case 0x32BD: retval = 48; break; // CIRCLED NUMBER FORTY EIGHT
644 case 0x32BE: retval = 49; break; // CIRCLED NUMBER FORTY NINE
645 case 0x32BF: retval = 50; break; // CIRCLED NUMBER FIFTY
646 default: retval = -2; break;
647 }
648 break;
649 case ($$valueJavaSupradecimal): // Java supradecimal
650 retval = (ch + ((val & $$maskDigitOffset) >> $$shiftDigitOffset) & $$maskDigit) + 10;
651 break;
652 }
653 return retval;
654 }
655
656 boolean isWhitespace(int ch) {
657 int props = getProperties(ch);
658 return ((props & $$maskIdentifierInfo) == $$valueJavaWhitespace);
659 }
660
661 byte getDirectionality(int ch) {
662 int val = getProperties(ch);
663 byte directionality = (byte)((val & $$maskBidi) >> $$shiftBidi);
664 if (directionality == 0xF ) {
665 switch(ch) {
666 case 0x202A :
667 // This is the only char with LRE
668 directionality = Character.DIRECTIONALITY_LEFT_TO_RIGHT_EMBEDDING;
669 break;
670 case 0x202B :
671 // This is the only char with RLE
672 directionality = Character.DIRECTIONALITY_RIGHT_TO_LEFT_EMBEDDING;
673 break;
674 case 0x202C :
675 // This is the only char with PDF
676 directionality = Character.DIRECTIONALITY_POP_DIRECTIONAL_FORMAT;
677 break;
678 case 0x202D :
679 // This is the only char with LRO
680 directionality = Character.DIRECTIONALITY_LEFT_TO_RIGHT_OVERRIDE;
681 break;
682 case 0x202E :
683 // This is the only char with RLO
684 directionality = Character.DIRECTIONALITY_RIGHT_TO_LEFT_OVERRIDE;
685 break;
686 case 0x2066 :
687 // This is the only char with LRI
688 directionality = Character.DIRECTIONALITY_LEFT_TO_RIGHT_ISOLATE;
689 break;
690 case 0x2067 :
691 // This is the only char with RLI
692 directionality = Character.DIRECTIONALITY_RIGHT_TO_LEFT_ISOLATE;
693 break;
694 case 0x2068 :
695 // This is the only char with FSI
696 directionality = Character.DIRECTIONALITY_FIRST_STRONG_ISOLATE;
697 break;
698 case 0x2069 :
699 // This is the only char with PDI
700 directionality = Character.DIRECTIONALITY_POP_DIRECTIONAL_ISOLATE;
701 break;
702 default :
703 directionality = Character.DIRECTIONALITY_UNDEFINED;
704 break;
705 }
706 }
707 return directionality;
708 }
709
710 boolean isMirrored(int ch) {
711 int props = getProperties(ch);
712 return ((props & $$maskMirrored) != 0);
713 }
714
715 int toUpperCaseEx(int ch) {
716 int mapChar = ch;
717 int val = getProperties(ch);
718
719 if ((val & $$maskUpperCase) != 0) {
720 if ((val & $$maskCaseOffset) != $$maskCaseOffset) {
721 int offset = val << $$shiftCaseOffsetSign >> ($$shiftCaseOffsetSign+$$shiftCaseOffset);
722 mapChar = ch - offset;
723 }
724 else {
725 switch(ch) {
726 case 0x017F: mapChar = 0x0053; break;
727 case 0x023F: mapChar = 0x2C7E; break;
728 case 0x0240: mapChar = 0x2C7F; break;
729 case 0x0250: mapChar = 0x2C6F; break;
730 case 0x0251: mapChar = 0x2C6D; break;
731 case 0x0252: mapChar = 0x2C70; break;
732 case 0x025C: mapChar = 0xA7AB; break;
733 case 0x0261: mapChar = 0xA7AC; break;
734 case 0x0265: mapChar = 0xA78D; break;
735 case 0x0266: mapChar = 0xA7AA; break;
736 case 0x026A: mapChar = 0xA7AE; break;
737 case 0x026B: mapChar = 0x2C62; break;
738 case 0x026C: mapChar = 0xA7AD; break;
739 case 0x0271: mapChar = 0x2C6E; break;
740 case 0x027D: mapChar = 0x2C64; break;
741 case 0x0287: mapChar = 0xA7B1; break;
742 case 0x029D: mapChar = 0xA7B2; break;
743 case 0x029E: mapChar = 0xA7B0; break;
744 case 0x1C80: mapChar = 0x0412; break;
745 case 0x1C81: mapChar = 0x0414; break;
746 case 0x1C82: mapChar = 0x041E; break;
747 case 0x1C83: mapChar = 0x0421; break;
748 case 0x1C84: mapChar = 0x0422; break;
749 case 0x1C85: mapChar = 0x0422; break;
750 case 0x1C86: mapChar = 0x042A; break;
751 case 0x1C87: mapChar = 0x0462; break;
752 case 0x1C88: mapChar = 0xA64A; break;
753 case 0x1D79: mapChar = 0xA77D; break;
754 case 0x1D7D: mapChar = 0x2C63; break;
755 case 0x1FBE: mapChar = 0x0399; break;
756 case 0x2C65: mapChar = 0x023A; break;
757 case 0x2C66: mapChar = 0x023E; break;
758 case 0x2D00: mapChar = 0x10A0; break;
759 case 0x2D01: mapChar = 0x10A1; break;
760 case 0x2D02: mapChar = 0x10A2; break;
761 case 0x2D03: mapChar = 0x10A3; break;
762 case 0x2D04: mapChar = 0x10A4; break;
763 case 0x2D05: mapChar = 0x10A5; break;
764 case 0x2D06: mapChar = 0x10A6; break;
765 case 0x2D07: mapChar = 0x10A7; break;
766 case 0x2D08: mapChar = 0x10A8; break;
767 case 0x2D09: mapChar = 0x10A9; break;
768 case 0x2D0A: mapChar = 0x10AA; break;
769 case 0x2D0B: mapChar = 0x10AB; break;
770 case 0x2D0C: mapChar = 0x10AC; break;
771 case 0x2D0D: mapChar = 0x10AD; break;
772 case 0x2D0E: mapChar = 0x10AE; break;
773 case 0x2D0F: mapChar = 0x10AF; break;
774 case 0x2D10: mapChar = 0x10B0; break;
775 case 0x2D11: mapChar = 0x10B1; break;
776 case 0x2D12: mapChar = 0x10B2; break;
777 case 0x2D13: mapChar = 0x10B3; break;
778 case 0x2D14: mapChar = 0x10B4; break;
779 case 0x2D15: mapChar = 0x10B5; break;
780 case 0x2D16: mapChar = 0x10B6; break;
781 case 0x2D17: mapChar = 0x10B7; break;
782 case 0x2D18: mapChar = 0x10B8; break;
783 case 0x2D19: mapChar = 0x10B9; break;
784 case 0x2D1A: mapChar = 0x10BA; break;
785 case 0x2D1B: mapChar = 0x10BB; break;
786 case 0x2D1C: mapChar = 0x10BC; break;
787 case 0x2D1D: mapChar = 0x10BD; break;
788 case 0x2D1E: mapChar = 0x10BE; break;
789 case 0x2D1F: mapChar = 0x10BF; break;
790 case 0x2D20: mapChar = 0x10C0; break;
791 case 0x2D21: mapChar = 0x10C1; break;
792 case 0x2D22: mapChar = 0x10C2; break;
793 case 0x2D23: mapChar = 0x10C3; break;
794 case 0x2D24: mapChar = 0x10C4; break;
795 case 0x2D25: mapChar = 0x10C5; break;
796 case 0x2D27: mapChar = 0x10C7; break;
797 case 0x2D2D: mapChar = 0x10CD; break;
798 case 0xAB53: mapChar = 0xA7B3; break;
799 case 0xAB70: mapChar = 0x13A0; break;
800 case 0xAB71: mapChar = 0x13A1; break;
801 case 0xAB72: mapChar = 0x13A2; break;
802 case 0xAB73: mapChar = 0x13A3; break;
803 case 0xAB74: mapChar = 0x13A4; break;
804 case 0xAB75: mapChar = 0x13A5; break;
805 case 0xAB76: mapChar = 0x13A6; break;
806 case 0xAB77: mapChar = 0x13A7; break;
807 case 0xAB78: mapChar = 0x13A8; break;
808 case 0xAB79: mapChar = 0x13A9; break;
809 case 0xAB7A: mapChar = 0x13AA; break;
810 case 0xAB7B: mapChar = 0x13AB; break;
811 case 0xAB7C: mapChar = 0x13AC; break;
812 case 0xAB7D: mapChar = 0x13AD; break;
813 case 0xAB7E: mapChar = 0x13AE; break;
814 case 0xAB7F: mapChar = 0x13AF; break;
815 case 0xAB80: mapChar = 0x13B0; break;
816 case 0xAB81: mapChar = 0x13B1; break;
817 case 0xAB82: mapChar = 0x13B2; break;
818 case 0xAB83: mapChar = 0x13B3; break;
819 case 0xAB84: mapChar = 0x13B4; break;
820 case 0xAB85: mapChar = 0x13B5; break;
821 case 0xAB86: mapChar = 0x13B6; break;
822 case 0xAB87: mapChar = 0x13B7; break;
823 case 0xAB88: mapChar = 0x13B8; break;
824 case 0xAB89: mapChar = 0x13B9; break;
825 case 0xAB8A: mapChar = 0x13BA; break;
826 case 0xAB8B: mapChar = 0x13BB; break;
827 case 0xAB8C: mapChar = 0x13BC; break;
828 case 0xAB8D: mapChar = 0x13BD; break;
829 case 0xAB8E: mapChar = 0x13BE; break;
830 case 0xAB8F: mapChar = 0x13BF; break;
831 case 0xAB90: mapChar = 0x13C0; break;
832 case 0xAB91: mapChar = 0x13C1; break;
833 case 0xAB92: mapChar = 0x13C2; break;
834 case 0xAB93: mapChar = 0x13C3; break;
835 case 0xAB94: mapChar = 0x13C4; break;
836 case 0xAB95: mapChar = 0x13C5; break;
837 case 0xAB96: mapChar = 0x13C6; break;
838 case 0xAB97: mapChar = 0x13C7; break;
839 case 0xAB98: mapChar = 0x13C8; break;
840 case 0xAB99: mapChar = 0x13C9; break;
841 case 0xAB9A: mapChar = 0x13CA; break;
842 case 0xAB9B: mapChar = 0x13CB; break;
843 case 0xAB9C: mapChar = 0x13CC; break;
844 case 0xAB9D: mapChar = 0x13CD; break;
845 case 0xAB9E: mapChar = 0x13CE; break;
846 case 0xAB9F: mapChar = 0x13CF; break;
847 case 0xABA0: mapChar = 0x13D0; break;
848 case 0xABA1: mapChar = 0x13D1; break;
849 case 0xABA2: mapChar = 0x13D2; break;
850 case 0xABA3: mapChar = 0x13D3; break;
851 case 0xABA4: mapChar = 0x13D4; break;
852 case 0xABA5: mapChar = 0x13D5; break;
853 case 0xABA6: mapChar = 0x13D6; break;
854 case 0xABA7: mapChar = 0x13D7; break;
855 case 0xABA8: mapChar = 0x13D8; break;
856 case 0xABA9: mapChar = 0x13D9; break;
857 case 0xABAA: mapChar = 0x13DA; break;
858 case 0xABAB: mapChar = 0x13DB; break;
859 case 0xABAC: mapChar = 0x13DC; break;
860 case 0xABAD: mapChar = 0x13DD; break;
861 case 0xABAE: mapChar = 0x13DE; break;
862 case 0xABAF: mapChar = 0x13DF; break;
863 case 0xABB0: mapChar = 0x13E0; break;
864 case 0xABB1: mapChar = 0x13E1; break;
865 case 0xABB2: mapChar = 0x13E2; break;
866 case 0xABB3: mapChar = 0x13E3; break;
867 case 0xABB4: mapChar = 0x13E4; break;
868 case 0xABB5: mapChar = 0x13E5; break;
869 case 0xABB6: mapChar = 0x13E6; break;
870 case 0xABB7: mapChar = 0x13E7; break;
871 case 0xABB8: mapChar = 0x13E8; break;
872 case 0xABB9: mapChar = 0x13E9; break;
873 case 0xABBA: mapChar = 0x13EA; break;
874 case 0xABBB: mapChar = 0x13EB; break;
875 case 0xABBC: mapChar = 0x13EC; break;
876 case 0xABBD: mapChar = 0x13ED; break;
877 case 0xABBE: mapChar = 0x13EE; break;
878 case 0xABBF: mapChar = 0x13EF; break;
879 default : mapChar = Character.ERROR; break;
880 }
881 }
882 }
883 return mapChar;
884 }
885
886 char[] toUpperCaseCharArray(int ch) {
887 char[] upperMap = {(char)ch};
888 int location = findInCharMap(ch);
889 if (location != -1) {
890 upperMap = charMap[location][1];
891 }
892 return upperMap;
893 }
894
895
896 /**
897 * Finds the character in the uppercase mapping table.
898 *
899 * @param ch the <code>char</code> to search
900 * @return the index location ch in the table or -1 if not found
901 * @since 1.4
902 */
903 int findInCharMap(int ch) {
904 if (charMap == null || charMap.length == 0) {
905 return -1;
906 }
907 int top, bottom, current;
908 bottom = 0;
909 top = charMap.length;
910 current = top/2;
911 // invariant: top > current >= bottom && ch >= CharacterData.charMap[bottom][0]
912 while (top - bottom > 1) {
913 if (ch >= charMap[current][0][0]) {
914 bottom = current;
915 } else {
916 top = current;
917 }
918 current = (top + bottom) / 2;
919 }
920 if (ch == charMap[current][0][0]) return current;
921 else return -1;
922 }
923
924 static final CharacterData00 instance = new CharacterData00();
925 private CharacterData00() {};
926
927 $$Tables
928
929 static {
930 $$Initializers
931 }
932 }