1 /*
2 * Copyright (c) 2000, 2017, 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.awt.font;
27
28 import java.io.IOException;
29 import java.io.ObjectOutputStream;
30 import java.util.Arrays;
31 import java.util.Comparator;
32 import java.util.EnumSet;
33 import java.util.Set;
34 import jdk.internal.misc.SharedSecrets;
35
36 /**
37 * The {@code NumericShaper} class is used to convert Latin-1 (European)
38 * digits to other Unicode decimal digits. Users of this class will
39 * primarily be people who wish to present data using
40 * national digit shapes, but find it more convenient to represent the
41 * data internally using Latin-1 (European) digits. This does not
42 * interpret the deprecated numeric shape selector character (U+206E).
43 * <p>
44 * Instances of {@code NumericShaper} are typically applied
45 * as attributes to text with the
46 * {@link TextAttribute#NUMERIC_SHAPING NUMERIC_SHAPING} attribute
47 * of the {@code TextAttribute} class.
48 * For example, this code snippet causes a {@code TextLayout} to
49 * shape European digits to Arabic in an Arabic context:<br>
50 * <blockquote><pre>
51 * Map map = new HashMap();
52 * map.put(TextAttribute.NUMERIC_SHAPING,
53 * NumericShaper.getContextualShaper(NumericShaper.ARABIC));
54 * FontRenderContext frc = ...;
55 * TextLayout layout = new TextLayout(text, map, frc);
56 * layout.draw(g2d, x, y);
57 * </pre></blockquote>
58 * <br>
59 * It is also possible to perform numeric shaping explicitly using instances
60 * of {@code NumericShaper}, as this code snippet demonstrates:<br>
61 * <blockquote><pre>
62 * char[] text = ...;
63 * // shape all EUROPEAN digits (except zero) to ARABIC digits
64 * NumericShaper shaper = NumericShaper.getShaper(NumericShaper.ARABIC);
65 * shaper.shape(text, start, count);
66 *
67 * // shape European digits to ARABIC digits if preceding text is Arabic, or
68 * // shape European digits to TAMIL digits if preceding text is Tamil, or
69 * // leave European digits alone if there is no preceding text, or
70 * // preceding text is neither Arabic nor Tamil
71 * NumericShaper shaper =
72 * NumericShaper.getContextualShaper(NumericShaper.ARABIC |
73 * NumericShaper.TAMIL,
74 * NumericShaper.EUROPEAN);
75 * shaper.shape(text, start, count);
76 * </pre></blockquote>
77 *
78 * <p><b>Bit mask- and enum-based Unicode ranges</b></p>
79 *
80 * <p>This class supports two different programming interfaces to
81 * represent Unicode ranges for script-specific digits: bit
82 * mask-based ones, such as {@link #ARABIC NumericShaper.ARABIC}, and
83 * enum-based ones, such as {@link NumericShaper.Range#ARABIC}.
84 * Multiple ranges can be specified by ORing bit mask-based constants,
85 * such as:
86 * <blockquote><pre>
87 * NumericShaper.ARABIC | NumericShaper.TAMIL
88 * </pre></blockquote>
89 * or creating a {@code Set} with the {@link NumericShaper.Range}
90 * constants, such as:
91 * <blockquote><pre>
92 * EnumSet.of(NumericShaper.Range.ARABIC, NumericShaper.Range.TAMIL)
93 * </pre></blockquote>
94 * The enum-based ranges are a super set of the bit mask-based ones.
95 *
96 * <p>If the two interfaces are mixed (including serialization),
97 * Unicode range values are mapped to their counterparts where such
98 * mapping is possible, such as {@code NumericShaper.Range.ARABIC}
99 * from/to {@code NumericShaper.ARABIC}. If any unmappable range
100 * values are specified, such as {@code NumericShaper.Range.BALINESE},
101 * those ranges are ignored.
102 *
103 * <p><b>Decimal Digits Precedence</b></p>
104 *
105 * <p>A Unicode range may have more than one set of decimal digits. If
106 * multiple decimal digits sets are specified for the same Unicode
107 * range, one of the sets will take precedence as follows.
108 *
109 * <table class="plain">
110 * <caption>NumericShaper constants precedence</caption>
111 * <thead>
112 * <tr>
113 * <th scope="col">Unicode Range
114 * <th scope="col">{@code NumericShaper} Constants
115 * <th scope="col">Precedence
116 * </thead>
117 * <tbody>
118 * <tr>
119 * <th scope="row" rowspan="2">Arabic
120 * <td>{@link NumericShaper#ARABIC NumericShaper.ARABIC}
121 * <br>
122 * {@link NumericShaper#EASTERN_ARABIC NumericShaper.EASTERN_ARABIC}
123 * <td>{@link NumericShaper#EASTERN_ARABIC NumericShaper.EASTERN_ARABIC}
124 * </tr>
125 * <tr>
126 * <th scope="row">{@link NumericShaper.Range#ARABIC}
127 * <br>
128 * {@link NumericShaper.Range#EASTERN_ARABIC}
129 * <td>{@link NumericShaper.Range#EASTERN_ARABIC}
130 * <tr>
131 * <th scope="row">Tai Tham
132 * <td>{@link NumericShaper.Range#TAI_THAM_HORA}
133 * <br>
134 * {@link NumericShaper.Range#TAI_THAM_THAM}
135 * <td>{@link NumericShaper.Range#TAI_THAM_THAM}
136 * </tbody>
137 * </table>
138 *
139 * @since 1.4
140 */
141 public final class NumericShaper implements java.io.Serializable {
142
143 // For access from java.text.Bidi
144 static {
145 if (SharedSecrets.getJavaAWTFontAccess() == null) {
146 SharedSecrets.setJavaAWTFontAccess(new JavaAWTFontAccessImpl());
147 }
148 }
149
150 /**
151 * A {@code NumericShaper.Range} represents a Unicode range of a
152 * script having its own decimal digits. For example, the {@link
153 * NumericShaper.Range#THAI} range has the Thai digits, THAI DIGIT
154 * ZERO (U+0E50) to THAI DIGIT NINE (U+0E59).
155 *
156 * <p>The {@code Range} enum replaces the traditional bit
157 * mask-based values (e.g., {@link NumericShaper#ARABIC}), and
158 * supports more Unicode ranges than the bit mask-based ones. For
159 * example, the following code using the bit mask:
160 * <blockquote><pre>
161 * NumericShaper.getContextualShaper(NumericShaper.ARABIC |
162 * NumericShaper.TAMIL,
163 * NumericShaper.EUROPEAN);
164 * </pre></blockquote>
165 * can be written using this enum as:
166 * <blockquote><pre>
167 * NumericShaper.getContextualShaper(EnumSet.of(
168 * NumericShaper.Range.ARABIC,
169 * NumericShaper.Range.TAMIL),
170 * NumericShaper.Range.EUROPEAN);
171 * </pre></blockquote>
172 *
173 * @since 1.7
174 */
175 public static enum Range {
176 // The order of EUROPEAN to MOGOLIAN must be consistent
177 // with the bitmask-based constants.
178 /**
179 * The Latin (European) range with the Latin (ASCII) digits.
180 */
181 EUROPEAN ('\u0030', '\u0000', '\u0300'),
182 /**
183 * The Arabic range with the Arabic-Indic digits.
184 */
185 ARABIC ('\u0660', '\u0600', '\u0780'),
186 /**
187 * The Arabic range with the Eastern Arabic-Indic digits.
188 */
189 EASTERN_ARABIC ('\u06f0', '\u0600', '\u0780'),
190 /**
191 * The Devanagari range with the Devanagari digits.
192 */
193 DEVANAGARI ('\u0966', '\u0900', '\u0980'),
194 /**
195 * The Bengali range with the Bengali digits.
196 */
197 BENGALI ('\u09e6', '\u0980', '\u0a00'),
198 /**
199 * The Gurmukhi range with the Gurmukhi digits.
200 */
201 GURMUKHI ('\u0a66', '\u0a00', '\u0a80'),
202 /**
203 * The Gujarati range with the Gujarati digits.
204 */
205 GUJARATI ('\u0ae6', '\u0b00', '\u0b80'),
206 /**
207 * The Oriya range with the Oriya digits.
208 */
209 ORIYA ('\u0b66', '\u0b00', '\u0b80'),
210 /**
211 * The Tamil range with the Tamil digits.
212 */
213 TAMIL ('\u0be6', '\u0b80', '\u0c00'),
214 /**
215 * The Telugu range with the Telugu digits.
216 */
217 TELUGU ('\u0c66', '\u0c00', '\u0c80'),
218 /**
219 * The Kannada range with the Kannada digits.
220 */
221 KANNADA ('\u0ce6', '\u0c80', '\u0d00'),
222 /**
223 * The Malayalam range with the Malayalam digits.
224 */
225 MALAYALAM ('\u0d66', '\u0d00', '\u0d80'),
226 /**
227 * The Thai range with the Thai digits.
228 */
229 THAI ('\u0e50', '\u0e00', '\u0e80'),
230 /**
231 * The Lao range with the Lao digits.
232 */
233 LAO ('\u0ed0', '\u0e80', '\u0f00'),
234 /**
235 * The Tibetan range with the Tibetan digits.
236 */
237 TIBETAN ('\u0f20', '\u0f00', '\u1000'),
238 /**
239 * The Myanmar range with the Myanmar digits.
240 */
241 MYANMAR ('\u1040', '\u1000', '\u1080'),
242 /**
243 * The Ethiopic range with the Ethiopic digits. Ethiopic
244 * does not have a decimal digit 0 so Latin (European) 0 is
245 * used.
246 */
247 ETHIOPIC ('\u1369', '\u1200', '\u1380') {
248 @Override
249 char getNumericBase() { return 1; }
250 },
251 /**
252 * The Khmer range with the Khmer digits.
253 */
254 KHMER ('\u17e0', '\u1780', '\u1800'),
255 /**
256 * The Mongolian range with the Mongolian digits.
257 */
258 MONGOLIAN ('\u1810', '\u1800', '\u1900'),
259 // The order of EUROPEAN to MOGOLIAN must be consistent
260 // with the bitmask-based constants.
261
262 /**
263 * The N'Ko range with the N'Ko digits.
264 */
265 NKO ('\u07c0', '\u07c0', '\u0800'),
266 /**
267 * The Myanmar range with the Myanmar Shan digits.
268 */
269 MYANMAR_SHAN ('\u1090', '\u1000', '\u10a0'),
270 /**
271 * The Limbu range with the Limbu digits.
272 */
273 LIMBU ('\u1946', '\u1900', '\u1950'),
274 /**
275 * The New Tai Lue range with the New Tai Lue digits.
276 */
277 NEW_TAI_LUE ('\u19d0', '\u1980', '\u19e0'),
278 /**
279 * The Balinese range with the Balinese digits.
280 */
281 BALINESE ('\u1b50', '\u1b00', '\u1b80'),
282 /**
283 * The Sundanese range with the Sundanese digits.
284 */
285 SUNDANESE ('\u1bb0', '\u1b80', '\u1bc0'),
286 /**
287 * The Lepcha range with the Lepcha digits.
288 */
289 LEPCHA ('\u1c40', '\u1c00', '\u1c50'),
290 /**
291 * The Ol Chiki range with the Ol Chiki digits.
292 */
293 OL_CHIKI ('\u1c50', '\u1c50', '\u1c80'),
294 /**
295 * The Vai range with the Vai digits.
296 */
297 VAI ('\ua620', '\ua500', '\ua640'),
298 /**
299 * The Saurashtra range with the Saurashtra digits.
300 */
301 SAURASHTRA ('\ua8d0', '\ua880', '\ua8e0'),
302 /**
303 * The Kayah Li range with the Kayah Li digits.
304 */
305 KAYAH_LI ('\ua900', '\ua900', '\ua930'),
306 /**
307 * The Cham range with the Cham digits.
308 */
309 CHAM ('\uaa50', '\uaa00', '\uaa60'),
310 /**
311 * The Tai Tham Hora range with the Tai Tham Hora digits.
312 */
313 TAI_THAM_HORA ('\u1a80', '\u1a20', '\u1ab0'),
314 /**
315 * The Tai Tham Tham range with the Tai Tham Tham digits.
316 */
317 TAI_THAM_THAM ('\u1a90', '\u1a20', '\u1ab0'),
318 /**
319 * The Javanese range with the Javanese digits.
320 */
321 JAVANESE ('\ua9d0', '\ua980', '\ua9e0'),
322 /**
323 * The Meetei Mayek range with the Meetei Mayek digits.
324 */
325 MEETEI_MAYEK ('\uabf0', '\uabc0', '\uac00'),
326 /**
327 * The Sinhala range with the Sinhala digits.
328 * @since 9
329 */
330 SINHALA ('\u0de6', '\u0d80', '\u0e00'),
331 /**
332 * The Myanmar Extended-B range with the Myanmar Tai Laing digits.
333 * @since 9
334 */
335 MYANMAR_TAI_LAING ('\ua9f0', '\ua9e0', '\uaa00');
336
337 private static int toRangeIndex(Range script) {
338 int index = script.ordinal();
339 return index < NUM_KEYS ? index : -1;
340 }
341
342 private static Range indexToRange(int index) {
343 return index < NUM_KEYS ? Range.values()[index] : null;
344 }
345
346 private static int toRangeMask(Set<Range> ranges) {
347 int m = 0;
348 for (Range range : ranges) {
349 int index = range.ordinal();
350 if (index < NUM_KEYS) {
351 m |= 1 << index;
352 }
353 }
354 return m;
355 }
356
357 private static Set<Range> maskToRangeSet(int mask) {
358 Set<Range> set = EnumSet.noneOf(Range.class);
359 Range[] a = Range.values();
360 for (int i = 0; i < NUM_KEYS; i++) {
361 if ((mask & (1 << i)) != 0) {
362 set.add(a[i]);
363 }
364 }
365 return set;
366 }
367
368 // base character of range digits
369 private final int base;
370 // Unicode range
371 private final int start, // inclusive
372 end; // exclusive
373
374 private Range(int base, int start, int end) {
375 this.base = base - ('0' + getNumericBase());
376 this.start = start;
377 this.end = end;
378 }
379
380 private int getDigitBase() {
381 return base;
382 }
383
384 char getNumericBase() {
385 return 0;
386 }
387
388 private boolean inRange(int c) {
389 return start <= c && c < end;
390 }
391 }
392
393 /** index of context for contextual shaping - values range from 0 to 18 */
394 private int key;
395
396 /** flag indicating whether to shape contextually (high bit) and which
397 * digit ranges to shape (bits 0-18)
398 */
399 private int mask;
400
401 /**
402 * The context {@code Range} for contextual shaping or the {@code
403 * Range} for non-contextual shaping. {@code null} for the bit
404 * mask-based API.
405 *
406 * @since 1.7
407 */
408 private Range shapingRange;
409
410 /**
411 * {@code Set<Range>} indicating which Unicode ranges to
412 * shape. {@code null} for the bit mask-based API.
413 */
414 private transient Set<Range> rangeSet;
415
416 /**
417 * rangeSet.toArray() value. Sorted by Range.base when the number
418 * of elements is greater than BSEARCH_THRESHOLD.
419 */
420 private transient Range[] rangeArray;
421
422 /**
423 * If more than BSEARCH_THRESHOLD ranges are specified, binary search is used.
424 */
425 private static final int BSEARCH_THRESHOLD = 3;
426
427 private static final long serialVersionUID = -8022764705923730308L;
428
429 /** Identifies the Latin-1 (European) and extended range, and
430 * Latin-1 (European) decimal base.
431 */
432 public static final int EUROPEAN = 1<<0;
433
434 /** Identifies the ARABIC range and decimal base. */
435 public static final int ARABIC = 1<<1;
436
437 /** Identifies the ARABIC range and ARABIC_EXTENDED decimal base. */
438 public static final int EASTERN_ARABIC = 1<<2;
439
440 /** Identifies the DEVANAGARI range and decimal base. */
441 public static final int DEVANAGARI = 1<<3;
442
443 /** Identifies the BENGALI range and decimal base. */
444 public static final int BENGALI = 1<<4;
445
446 /** Identifies the GURMUKHI range and decimal base. */
447 public static final int GURMUKHI = 1<<5;
448
449 /** Identifies the GUJARATI range and decimal base. */
450 public static final int GUJARATI = 1<<6;
451
452 /** Identifies the ORIYA range and decimal base. */
453 public static final int ORIYA = 1<<7;
454
455 /** Identifies the TAMIL range and decimal base. */
456 // TAMIL DIGIT ZERO was added in Unicode 4.1
457 public static final int TAMIL = 1<<8;
458
459 /** Identifies the TELUGU range and decimal base. */
460 public static final int TELUGU = 1<<9;
461
462 /** Identifies the KANNADA range and decimal base. */
463 public static final int KANNADA = 1<<10;
464
465 /** Identifies the MALAYALAM range and decimal base. */
466 public static final int MALAYALAM = 1<<11;
467
468 /** Identifies the THAI range and decimal base. */
469 public static final int THAI = 1<<12;
470
471 /** Identifies the LAO range and decimal base. */
472 public static final int LAO = 1<<13;
473
474 /** Identifies the TIBETAN range and decimal base. */
475 public static final int TIBETAN = 1<<14;
476
477 /** Identifies the MYANMAR range and decimal base. */
478 public static final int MYANMAR = 1<<15;
479
480 /** Identifies the ETHIOPIC range and decimal base. */
481 public static final int ETHIOPIC = 1<<16;
482
483 /** Identifies the KHMER range and decimal base. */
484 public static final int KHMER = 1<<17;
485
486 /** Identifies the MONGOLIAN range and decimal base. */
487 public static final int MONGOLIAN = 1<<18;
488
489 /** Identifies all ranges, for full contextual shaping.
490 *
491 * <p>This constant specifies all of the bit mask-based
492 * ranges. Use {@code EnumSet.allOf(NumericShaper.Range.class)} to
493 * specify all of the enum-based ranges.
494 */
495 public static final int ALL_RANGES = 0x0007ffff;
496
497 private static final int EUROPEAN_KEY = 0;
498 private static final int ARABIC_KEY = 1;
499 private static final int EASTERN_ARABIC_KEY = 2;
500 private static final int DEVANAGARI_KEY = 3;
501 private static final int BENGALI_KEY = 4;
502 private static final int GURMUKHI_KEY = 5;
503 private static final int GUJARATI_KEY = 6;
504 private static final int ORIYA_KEY = 7;
505 private static final int TAMIL_KEY = 8;
506 private static final int TELUGU_KEY = 9;
507 private static final int KANNADA_KEY = 10;
508 private static final int MALAYALAM_KEY = 11;
509 private static final int THAI_KEY = 12;
510 private static final int LAO_KEY = 13;
511 private static final int TIBETAN_KEY = 14;
512 private static final int MYANMAR_KEY = 15;
513 private static final int ETHIOPIC_KEY = 16;
514 private static final int KHMER_KEY = 17;
515 private static final int MONGOLIAN_KEY = 18;
516
517 private static final int NUM_KEYS = MONGOLIAN_KEY + 1; // fixed
518
519 private static final int CONTEXTUAL_MASK = 1<<31;
520
521 private static final char[] bases = {
522 '\u0030' - '\u0030', // EUROPEAN
523 '\u0660' - '\u0030', // ARABIC-INDIC
524 '\u06f0' - '\u0030', // EXTENDED ARABIC-INDIC (EASTERN_ARABIC)
525 '\u0966' - '\u0030', // DEVANAGARI
526 '\u09e6' - '\u0030', // BENGALI
527 '\u0a66' - '\u0030', // GURMUKHI
528 '\u0ae6' - '\u0030', // GUJARATI
529 '\u0b66' - '\u0030', // ORIYA
530 '\u0be6' - '\u0030', // TAMIL - zero was added in Unicode 4.1
531 '\u0c66' - '\u0030', // TELUGU
532 '\u0ce6' - '\u0030', // KANNADA
533 '\u0d66' - '\u0030', // MALAYALAM
534 '\u0e50' - '\u0030', // THAI
535 '\u0ed0' - '\u0030', // LAO
536 '\u0f20' - '\u0030', // TIBETAN
537 '\u1040' - '\u0030', // MYANMAR
538 '\u1369' - '\u0031', // ETHIOPIC - no zero
539 '\u17e0' - '\u0030', // KHMER
540 '\u1810' - '\u0030', // MONGOLIAN
541 };
542
543 // some ranges adjoin or overlap, rethink if we want to do a binary search on this
544
545 private static final char[] contexts = {
546 '\u0000', '\u0300', // 'EUROPEAN' (really latin-1 and extended)
547 '\u0600', '\u0780', // ARABIC
548 '\u0600', '\u0780', // EASTERN_ARABIC -- note overlap with arabic
549 '\u0900', '\u0980', // DEVANAGARI
550 '\u0980', '\u0a00', // BENGALI
551 '\u0a00', '\u0a80', // GURMUKHI
552 '\u0a80', '\u0b00', // GUJARATI
553 '\u0b00', '\u0b80', // ORIYA
554 '\u0b80', '\u0c00', // TAMIL
555 '\u0c00', '\u0c80', // TELUGU
556 '\u0c80', '\u0d00', // KANNADA
557 '\u0d00', '\u0d80', // MALAYALAM
558 '\u0e00', '\u0e80', // THAI
559 '\u0e80', '\u0f00', // LAO
560 '\u0f00', '\u1000', // TIBETAN
561 '\u1000', '\u1080', // MYANMAR
562 '\u1200', '\u1380', // ETHIOPIC - note missing zero
563 '\u1780', '\u1800', // KHMER
564 '\u1800', '\u1900', // MONGOLIAN
565 '\uffff',
566 };
567
568 // assume most characters are near each other so probing the cache is infrequent,
569 // and a linear probe is ok.
570
571 private static int ctCache = 0;
572 private static int ctCacheLimit = contexts.length - 2;
573
574 // warning, synchronize access to this as it modifies state
575 private static int getContextKey(char c) {
576 if (c < contexts[ctCache]) {
577 while (ctCache > 0 && c < contexts[ctCache]) --ctCache;
578 } else if (c >= contexts[ctCache + 1]) {
579 while (ctCache < ctCacheLimit && c >= contexts[ctCache + 1]) ++ctCache;
580 }
581
582 // if we're not in a known range, then return EUROPEAN as the range key
583 return (ctCache & 0x1) == 0 ? (ctCache / 2) : EUROPEAN_KEY;
584 }
585
586 // cache for the NumericShaper.Range version
587 private transient volatile Range currentRange = Range.EUROPEAN;
588
589 private Range rangeForCodePoint(final int codepoint) {
590 if (currentRange.inRange(codepoint)) {
591 return currentRange;
592 }
593
594 final Range[] ranges = rangeArray;
595 if (ranges.length > BSEARCH_THRESHOLD) {
596 int lo = 0;
597 int hi = ranges.length - 1;
598 while (lo <= hi) {
599 int mid = (lo + hi) / 2;
600 Range range = ranges[mid];
601 if (codepoint < range.start) {
602 hi = mid - 1;
603 } else if (codepoint >= range.end) {
604 lo = mid + 1;
605 } else {
606 currentRange = range;
607 return range;
608 }
609 }
610 } else {
611 for (int i = 0; i < ranges.length; i++) {
612 if (ranges[i].inRange(codepoint)) {
613 return ranges[i];
614 }
615 }
616 }
617 return Range.EUROPEAN;
618 }
619
620 /*
621 * A range table of strong directional characters (types L, R, AL).
622 * Even (left) indexes are starts of ranges of non-strong-directional (or undefined)
623 * characters, odd (right) indexes are starts of ranges of strong directional
624 * characters.
625 */
626 private static int[] strongTable = {
627 0x0000, 0x0041,
628 0x005b, 0x0061,
629 0x007b, 0x00aa,
630 0x00ab, 0x00b5,
631 0x00b6, 0x00ba,
632 0x00bb, 0x00c0,
633 0x00d7, 0x00d8,
634 0x00f7, 0x00f8,
635 0x02b9, 0x02bb,
636 0x02c2, 0x02d0,
637 0x02d2, 0x02e0,
638 0x02e5, 0x02ee,
639 0x02ef, 0x0370,
640 0x0374, 0x0376,
641 0x0378, 0x037a,
642 0x037e, 0x037f,
643 0x0380, 0x0386,
644 0x0387, 0x0388,
645 0x038b, 0x038c,
646 0x038d, 0x038e,
647 0x03a2, 0x03a3,
648 0x03f6, 0x03f7,
649 0x0483, 0x048a,
650 0x0530, 0x0531,
651 0x0557, 0x0559,
652 0x0560, 0x0561,
653 0x0588, 0x0589,
654 0x058a, 0x0590,
655 0x0591, 0x05be,
656 0x05bf, 0x05c0,
657 0x05c1, 0x05c3,
658 0x05c4, 0x05c6,
659 0x05c7, 0x05c8,
660 0x0600, 0x0608,
661 0x0609, 0x060b,
662 0x060c, 0x060d,
663 0x060e, 0x061b,
664 0x064b, 0x066d,
665 0x0670, 0x0671,
666 0x06d6, 0x06e5,
667 0x06e7, 0x06ee,
668 0x06f0, 0x06fa,
669 0x0711, 0x0712,
670 0x0730, 0x074b,
671 0x07a6, 0x07b1,
672 0x07eb, 0x07f4,
673 0x07f6, 0x07fa,
674 0x0816, 0x081a,
675 0x081b, 0x0824,
676 0x0825, 0x0828,
677 0x0829, 0x082e,
678 0x0859, 0x085c,
679 0x08e3, 0x0903,
680 0x093a, 0x093b,
681 0x093c, 0x093d,
682 0x0941, 0x0949,
683 0x094d, 0x094e,
684 0x0951, 0x0958,
685 0x0962, 0x0964,
686 0x0981, 0x0982,
687 0x0984, 0x0985,
688 0x098d, 0x098f,
689 0x0991, 0x0993,
690 0x09a9, 0x09aa,
691 0x09b1, 0x09b2,
692 0x09b3, 0x09b6,
693 0x09ba, 0x09bd,
694 0x09c1, 0x09c7,
695 0x09c9, 0x09cb,
696 0x09cd, 0x09ce,
697 0x09cf, 0x09d7,
698 0x09d8, 0x09dc,
699 0x09de, 0x09df,
700 0x09e2, 0x09e6,
701 0x09f2, 0x09f4,
702 0x09fb, 0x0a03,
703 0x0a04, 0x0a05,
704 0x0a0b, 0x0a0f,
705 0x0a11, 0x0a13,
706 0x0a29, 0x0a2a,
707 0x0a31, 0x0a32,
708 0x0a34, 0x0a35,
709 0x0a37, 0x0a38,
710 0x0a3a, 0x0a3e,
711 0x0a41, 0x0a59,
712 0x0a5d, 0x0a5e,
713 0x0a5f, 0x0a66,
714 0x0a70, 0x0a72,
715 0x0a75, 0x0a83,
716 0x0a84, 0x0a85,
717 0x0a8e, 0x0a8f,
718 0x0a92, 0x0a93,
719 0x0aa9, 0x0aaa,
720 0x0ab1, 0x0ab2,
721 0x0ab4, 0x0ab5,
722 0x0aba, 0x0abd,
723 0x0ac1, 0x0ac9,
724 0x0aca, 0x0acb,
725 0x0acd, 0x0ad0,
726 0x0ad1, 0x0ae0,
727 0x0ae2, 0x0ae6,
728 0x0af1, 0x0af9,
729 0x0afa, 0x0b02,
730 0x0b04, 0x0b05,
731 0x0b0d, 0x0b0f,
732 0x0b11, 0x0b13,
733 0x0b29, 0x0b2a,
734 0x0b31, 0x0b32,
735 0x0b34, 0x0b35,
736 0x0b3a, 0x0b3d,
737 0x0b3f, 0x0b40,
738 0x0b41, 0x0b47,
739 0x0b49, 0x0b4b,
740 0x0b4d, 0x0b57,
741 0x0b58, 0x0b5c,
742 0x0b5e, 0x0b5f,
743 0x0b62, 0x0b66,
744 0x0b78, 0x0b83,
745 0x0b84, 0x0b85,
746 0x0b8b, 0x0b8e,
747 0x0b91, 0x0b92,
748 0x0b96, 0x0b99,
749 0x0b9b, 0x0b9c,
750 0x0b9d, 0x0b9e,
751 0x0ba0, 0x0ba3,
752 0x0ba5, 0x0ba8,
753 0x0bab, 0x0bae,
754 0x0bba, 0x0bbe,
755 0x0bc0, 0x0bc1,
756 0x0bc3, 0x0bc6,
757 0x0bc9, 0x0bca,
758 0x0bcd, 0x0bd0,
759 0x0bd1, 0x0bd7,
760 0x0bd8, 0x0be6,
761 0x0bf3, 0x0c01,
762 0x0c04, 0x0c05,
763 0x0c0d, 0x0c0e,
764 0x0c11, 0x0c12,
765 0x0c29, 0x0c2a,
766 0x0c3a, 0x0c3d,
767 0x0c3e, 0x0c41,
768 0x0c45, 0x0c58,
769 0x0c5b, 0x0c60,
770 0x0c62, 0x0c66,
771 0x0c70, 0x0c7f,
772 0x0c80, 0x0c82,
773 0x0c84, 0x0c85,
774 0x0c8d, 0x0c8e,
775 0x0c91, 0x0c92,
776 0x0ca9, 0x0caa,
777 0x0cb4, 0x0cb5,
778 0x0cba, 0x0cbd,
779 0x0cc5, 0x0cc6,
780 0x0cc9, 0x0cca,
781 0x0ccc, 0x0cd5,
782 0x0cd7, 0x0cde,
783 0x0cdf, 0x0ce0,
784 0x0ce2, 0x0ce6,
785 0x0cf0, 0x0cf1,
786 0x0cf3, 0x0d02,
787 0x0d04, 0x0d05,
788 0x0d0d, 0x0d0e,
789 0x0d11, 0x0d12,
790 0x0d3b, 0x0d3d,
791 0x0d41, 0x0d46,
792 0x0d49, 0x0d4a,
793 0x0d4d, 0x0d4e,
794 0x0d4f, 0x0d57,
795 0x0d58, 0x0d5f,
796 0x0d62, 0x0d66,
797 0x0d76, 0x0d79,
798 0x0d80, 0x0d82,
799 0x0d84, 0x0d85,
800 0x0d97, 0x0d9a,
801 0x0db2, 0x0db3,
802 0x0dbc, 0x0dbd,
803 0x0dbe, 0x0dc0,
804 0x0dc7, 0x0dcf,
805 0x0dd2, 0x0dd8,
806 0x0de0, 0x0de6,
807 0x0df0, 0x0df2,
808 0x0df5, 0x0e01,
809 0x0e31, 0x0e32,
810 0x0e34, 0x0e40,
811 0x0e47, 0x0e4f,
812 0x0e5c, 0x0e81,
813 0x0e83, 0x0e84,
814 0x0e85, 0x0e87,
815 0x0e89, 0x0e8a,
816 0x0e8b, 0x0e8d,
817 0x0e8e, 0x0e94,
818 0x0e98, 0x0e99,
819 0x0ea0, 0x0ea1,
820 0x0ea4, 0x0ea5,
821 0x0ea6, 0x0ea7,
822 0x0ea8, 0x0eaa,
823 0x0eac, 0x0ead,
824 0x0eb1, 0x0eb2,
825 0x0eb4, 0x0ebd,
826 0x0ebe, 0x0ec0,
827 0x0ec5, 0x0ec6,
828 0x0ec7, 0x0ed0,
829 0x0eda, 0x0edc,
830 0x0ee0, 0x0f00,
831 0x0f18, 0x0f1a,
832 0x0f35, 0x0f36,
833 0x0f37, 0x0f38,
834 0x0f39, 0x0f3e,
835 0x0f48, 0x0f49,
836 0x0f6d, 0x0f7f,
837 0x0f80, 0x0f85,
838 0x0f86, 0x0f88,
839 0x0f8d, 0x0fbe,
840 0x0fc6, 0x0fc7,
841 0x0fcd, 0x0fce,
842 0x0fdb, 0x1000,
843 0x102d, 0x1031,
844 0x1032, 0x1038,
845 0x1039, 0x103b,
846 0x103d, 0x103f,
847 0x1058, 0x105a,
848 0x105e, 0x1061,
849 0x1071, 0x1075,
850 0x1082, 0x1083,
851 0x1085, 0x1087,
852 0x108d, 0x108e,
853 0x109d, 0x109e,
854 0x10c6, 0x10c7,
855 0x10c8, 0x10cd,
856 0x10ce, 0x10d0,
857 0x1249, 0x124a,
858 0x124e, 0x1250,
859 0x1257, 0x1258,
860 0x1259, 0x125a,
861 0x125e, 0x1260,
862 0x1289, 0x128a,
863 0x128e, 0x1290,
864 0x12b1, 0x12b2,
865 0x12b6, 0x12b8,
866 0x12bf, 0x12c0,
867 0x12c1, 0x12c2,
868 0x12c6, 0x12c8,
869 0x12d7, 0x12d8,
870 0x1311, 0x1312,
871 0x1316, 0x1318,
872 0x135b, 0x1360,
873 0x137d, 0x1380,
874 0x1390, 0x13a0,
875 0x13f6, 0x13f8,
876 0x13fe, 0x1401,
877 0x1680, 0x1681,
878 0x169b, 0x16a0,
879 0x16f9, 0x1700,
880 0x170d, 0x170e,
881 0x1712, 0x1720,
882 0x1732, 0x1735,
883 0x1737, 0x1740,
884 0x1752, 0x1760,
885 0x176d, 0x176e,
886 0x1771, 0x1780,
887 0x17b4, 0x17b6,
888 0x17b7, 0x17be,
889 0x17c6, 0x17c7,
890 0x17c9, 0x17d4,
891 0x17db, 0x17dc,
892 0x17dd, 0x17e0,
893 0x17ea, 0x1810,
894 0x181a, 0x1820,
895 0x1878, 0x1880,
896 0x18a9, 0x18aa,
897 0x18ab, 0x18b0,
898 0x18f6, 0x1900,
899 0x191f, 0x1923,
900 0x1927, 0x1929,
901 0x192c, 0x1930,
902 0x1932, 0x1933,
903 0x1939, 0x1946,
904 0x196e, 0x1970,
905 0x1975, 0x1980,
906 0x19ac, 0x19b0,
907 0x19ca, 0x19d0,
908 0x19db, 0x1a00,
909 0x1a17, 0x1a19,
910 0x1a1b, 0x1a1e,
911 0x1a56, 0x1a57,
912 0x1a58, 0x1a61,
913 0x1a62, 0x1a63,
914 0x1a65, 0x1a6d,
915 0x1a73, 0x1a80,
916 0x1a8a, 0x1a90,
917 0x1a9a, 0x1aa0,
918 0x1aae, 0x1b04,
919 0x1b34, 0x1b35,
920 0x1b36, 0x1b3b,
921 0x1b3c, 0x1b3d,
922 0x1b42, 0x1b43,
923 0x1b4c, 0x1b50,
924 0x1b6b, 0x1b74,
925 0x1b7d, 0x1b82,
926 0x1ba2, 0x1ba6,
927 0x1ba8, 0x1baa,
928 0x1bab, 0x1bae,
929 0x1be6, 0x1be7,
930 0x1be8, 0x1bea,
931 0x1bed, 0x1bee,
932 0x1bef, 0x1bf2,
933 0x1bf4, 0x1bfc,
934 0x1c2c, 0x1c34,
935 0x1c36, 0x1c3b,
936 0x1c4a, 0x1c4d,
937 0x1c80, 0x1cc0,
938 0x1cc8, 0x1cd3,
939 0x1cd4, 0x1ce1,
940 0x1ce2, 0x1ce9,
941 0x1ced, 0x1cee,
942 0x1cf4, 0x1cf5,
943 0x1cf7, 0x1d00,
944 0x1dc0, 0x1e00,
945 0x1f16, 0x1f18,
946 0x1f1e, 0x1f20,
947 0x1f46, 0x1f48,
948 0x1f4e, 0x1f50,
949 0x1f58, 0x1f59,
950 0x1f5a, 0x1f5b,
951 0x1f5c, 0x1f5d,
952 0x1f5e, 0x1f5f,
953 0x1f7e, 0x1f80,
954 0x1fb5, 0x1fb6,
955 0x1fbd, 0x1fbe,
956 0x1fbf, 0x1fc2,
957 0x1fc5, 0x1fc6,
958 0x1fcd, 0x1fd0,
959 0x1fd4, 0x1fd6,
960 0x1fdc, 0x1fe0,
961 0x1fed, 0x1ff2,
962 0x1ff5, 0x1ff6,
963 0x1ffd, 0x200e,
964 0x2010, 0x2071,
965 0x2072, 0x207f,
966 0x2080, 0x2090,
967 0x209d, 0x2102,
968 0x2103, 0x2107,
969 0x2108, 0x210a,
970 0x2114, 0x2115,
971 0x2116, 0x2119,
972 0x211e, 0x2124,
973 0x2125, 0x2126,
974 0x2127, 0x2128,
975 0x2129, 0x212a,
976 0x212e, 0x212f,
977 0x213a, 0x213c,
978 0x2140, 0x2145,
979 0x214a, 0x214e,
980 0x2150, 0x2160,
981 0x2189, 0x2336,
982 0x237b, 0x2395,
983 0x2396, 0x249c,
984 0x24ea, 0x26ac,
985 0x26ad, 0x2800,
986 0x2900, 0x2c00,
987 0x2c2f, 0x2c30,
988 0x2c5f, 0x2c60,
989 0x2ce5, 0x2ceb,
990 0x2cef, 0x2cf2,
991 0x2cf4, 0x2d00,
992 0x2d26, 0x2d27,
993 0x2d28, 0x2d2d,
994 0x2d2e, 0x2d30,
995 0x2d68, 0x2d6f,
996 0x2d71, 0x2d80,
997 0x2d97, 0x2da0,
998 0x2da7, 0x2da8,
999 0x2daf, 0x2db0,
1000 0x2db7, 0x2db8,
1001 0x2dbf, 0x2dc0,
1002 0x2dc7, 0x2dc8,
1003 0x2dcf, 0x2dd0,
1004 0x2dd7, 0x2dd8,
1005 0x2ddf, 0x3005,
1006 0x3008, 0x3021,
1007 0x302a, 0x302e,
1008 0x3030, 0x3031,
1009 0x3036, 0x3038,
1010 0x303d, 0x3041,
1011 0x3097, 0x309d,
1012 0x30a0, 0x30a1,
1013 0x30fb, 0x30fc,
1014 0x3100, 0x3105,
1015 0x312e, 0x3131,
1016 0x318f, 0x3190,
1017 0x31bb, 0x31f0,
1018 0x321d, 0x3220,
1019 0x3250, 0x3260,
1020 0x327c, 0x327f,
1021 0x32b1, 0x32c0,
1022 0x32cc, 0x32d0,
1023 0x32ff, 0x3300,
1024 0x3377, 0x337b,
1025 0x33de, 0x33e0,
1026 0x33ff, 0x3400,
1027 0x4db6, 0x4e00,
1028 0x9fd6, 0xa000,
1029 0xa48d, 0xa4d0,
1030 0xa60d, 0xa610,
1031 0xa62c, 0xa640,
1032 0xa66f, 0xa680,
1033 0xa69e, 0xa6a0,
1034 0xa6f0, 0xa6f2,
1035 0xa6f8, 0xa722,
1036 0xa788, 0xa789,
1037 0xa7ae, 0xa7b0,
1038 0xa7b8, 0xa7f7,
1039 0xa802, 0xa803,
1040 0xa806, 0xa807,
1041 0xa80b, 0xa80c,
1042 0xa825, 0xa827,
1043 0xa828, 0xa830,
1044 0xa838, 0xa840,
1045 0xa874, 0xa880,
1046 0xa8c4, 0xa8ce,
1047 0xa8da, 0xa8f2,
1048 0xa8fe, 0xa900,
1049 0xa926, 0xa92e,
1050 0xa947, 0xa952,
1051 0xa954, 0xa95f,
1052 0xa97d, 0xa983,
1053 0xa9b3, 0xa9b4,
1054 0xa9b6, 0xa9ba,
1055 0xa9bc, 0xa9bd,
1056 0xa9ce, 0xa9cf,
1057 0xa9da, 0xa9de,
1058 0xa9e5, 0xa9e6,
1059 0xa9ff, 0xaa00,
1060 0xaa29, 0xaa2f,
1061 0xaa31, 0xaa33,
1062 0xaa35, 0xaa40,
1063 0xaa43, 0xaa44,
1064 0xaa4c, 0xaa4d,
1065 0xaa4e, 0xaa50,
1066 0xaa5a, 0xaa5c,
1067 0xaa7c, 0xaa7d,
1068 0xaab0, 0xaab1,
1069 0xaab2, 0xaab5,
1070 0xaab7, 0xaab9,
1071 0xaabe, 0xaac0,
1072 0xaac1, 0xaac2,
1073 0xaac3, 0xaadb,
1074 0xaaec, 0xaaee,
1075 0xaaf6, 0xab01,
1076 0xab07, 0xab09,
1077 0xab0f, 0xab11,
1078 0xab17, 0xab20,
1079 0xab27, 0xab28,
1080 0xab2f, 0xab30,
1081 0xab66, 0xab70,
1082 0xabe5, 0xabe6,
1083 0xabe8, 0xabe9,
1084 0xabed, 0xabf0,
1085 0xabfa, 0xac00,
1086 0xd7a4, 0xd7b0,
1087 0xd7c7, 0xd7cb,
1088 0xd7fc, 0xe000,
1089 0xfa6e, 0xfa70,
1090 0xfada, 0xfb00,
1091 0xfb07, 0xfb13,
1092 0xfb18, 0xfb1d,
1093 0xfb1e, 0xfb1f,
1094 0xfb29, 0xfb2a,
1095 0xfd3e, 0xfd40,
1096 0xfdd0, 0xfdf0,
1097 0xfdfd, 0xfdfe,
1098 0xfe00, 0xfe70,
1099 0xfeff, 0xff21,
1100 0xff3b, 0xff41,
1101 0xff5b, 0xff66,
1102 0xffbf, 0xffc2,
1103 0xffc8, 0xffca,
1104 0xffd0, 0xffd2,
1105 0xffd8, 0xffda,
1106 0xffdd, 0x10000,
1107 0x1000c, 0x1000d,
1108 0x10027, 0x10028,
1109 0x1003b, 0x1003c,
1110 0x1003e, 0x1003f,
1111 0x1004e, 0x10050,
1112 0x1005e, 0x10080,
1113 0x100fb, 0x10100,
1114 0x10101, 0x10102,
1115 0x10103, 0x10107,
1116 0x10134, 0x10137,
1117 0x10140, 0x101d0,
1118 0x101fd, 0x10280,
1119 0x1029d, 0x102a0,
1120 0x102d1, 0x10300,
1121 0x10324, 0x10330,
1122 0x1034b, 0x10350,
1123 0x10376, 0x10380,
1124 0x1039e, 0x1039f,
1125 0x103c4, 0x103c8,
1126 0x103d6, 0x10400,
1127 0x1049e, 0x104a0,
1128 0x104aa, 0x10500,
1129 0x10528, 0x10530,
1130 0x10564, 0x1056f,
1131 0x10570, 0x10600,
1132 0x10737, 0x10740,
1133 0x10756, 0x10760,
1134 0x10768, 0x10800,
1135 0x1091f, 0x10920,
1136 0x10a01, 0x10a04,
1137 0x10a05, 0x10a07,
1138 0x10a0c, 0x10a10,
1139 0x10a38, 0x10a3b,
1140 0x10a3f, 0x10a40,
1141 0x10ae5, 0x10ae7,
1142 0x10b39, 0x10b40,
1143 0x10e60, 0x10e7f,
1144 0x11001, 0x11002,
1145 0x11038, 0x11047,
1146 0x1104e, 0x11066,
1147 0x11070, 0x11082,
1148 0x110b3, 0x110b7,
1149 0x110b9, 0x110bb,
1150 0x110c2, 0x110d0,
1151 0x110e9, 0x110f0,
1152 0x110fa, 0x11103,
1153 0x11127, 0x1112c,
1154 0x1112d, 0x11136,
1155 0x11144, 0x11150,
1156 0x11173, 0x11174,
1157 0x11177, 0x11182,
1158 0x111b6, 0x111bf,
1159 0x111ca, 0x111cd,
1160 0x111ce, 0x111d0,
1161 0x111e0, 0x111e1,
1162 0x111f5, 0x11200,
1163 0x11212, 0x11213,
1164 0x1122f, 0x11232,
1165 0x11234, 0x11235,
1166 0x11236, 0x11238,
1167 0x1123e, 0x11280,
1168 0x11287, 0x11288,
1169 0x11289, 0x1128a,
1170 0x1128e, 0x1128f,
1171 0x1129e, 0x1129f,
1172 0x112aa, 0x112b0,
1173 0x112df, 0x112e0,
1174 0x112e3, 0x112f0,
1175 0x112fa, 0x11302,
1176 0x11304, 0x11305,
1177 0x1130d, 0x1130f,
1178 0x11311, 0x11313,
1179 0x11329, 0x1132a,
1180 0x11331, 0x11332,
1181 0x11334, 0x11335,
1182 0x1133a, 0x1133d,
1183 0x11340, 0x11341,
1184 0x11345, 0x11347,
1185 0x11349, 0x1134b,
1186 0x1134e, 0x11350,
1187 0x11351, 0x11357,
1188 0x11358, 0x1135d,
1189 0x11364, 0x11480,
1190 0x114b3, 0x114b9,
1191 0x114ba, 0x114bb,
1192 0x114bf, 0x114c1,
1193 0x114c2, 0x114c4,
1194 0x114c8, 0x114d0,
1195 0x114da, 0x11580,
1196 0x115b2, 0x115b8,
1197 0x115bc, 0x115be,
1198 0x115bf, 0x115c1,
1199 0x115dc, 0x11600,
1200 0x11633, 0x1163b,
1201 0x1163d, 0x1163e,
1202 0x1163f, 0x11641,
1203 0x11645, 0x11650,
1204 0x1165a, 0x11680,
1205 0x116ab, 0x116ac,
1206 0x116ad, 0x116ae,
1207 0x116b0, 0x116b6,
1208 0x116b7, 0x116c0,
1209 0x116ca, 0x11700,
1210 0x1171a, 0x11720,
1211 0x11722, 0x11726,
1212 0x11727, 0x11730,
1213 0x11740, 0x118a0,
1214 0x118f3, 0x118ff,
1215 0x11900, 0x11ac0,
1216 0x11af9, 0x12000,
1217 0x1239a, 0x12400,
1218 0x1246f, 0x12470,
1219 0x12475, 0x12480,
1220 0x12544, 0x13000,
1221 0x1342f, 0x14400,
1222 0x14647, 0x16800,
1223 0x16a39, 0x16a40,
1224 0x16a5f, 0x16a60,
1225 0x16a6a, 0x16a6e,
1226 0x16a70, 0x16ad0,
1227 0x16aee, 0x16af5,
1228 0x16af6, 0x16b00,
1229 0x16b30, 0x16b37,
1230 0x16b46, 0x16b50,
1231 0x16b5a, 0x16b5b,
1232 0x16b62, 0x16b63,
1233 0x16b78, 0x16b7d,
1234 0x16b90, 0x16f00,
1235 0x16f45, 0x16f50,
1236 0x16f7f, 0x16f93,
1237 0x16fa0, 0x1b000,
1238 0x1b002, 0x1bc00,
1239 0x1bc6b, 0x1bc70,
1240 0x1bc7d, 0x1bc80,
1241 0x1bc89, 0x1bc90,
1242 0x1bc9a, 0x1bc9c,
1243 0x1bc9d, 0x1bc9f,
1244 0x1bca0, 0x1d000,
1245 0x1d0f6, 0x1d100,
1246 0x1d127, 0x1d129,
1247 0x1d167, 0x1d16a,
1248 0x1d173, 0x1d183,
1249 0x1d185, 0x1d18c,
1250 0x1d1aa, 0x1d1ae,
1251 0x1d1e9, 0x1d360,
1252 0x1d372, 0x1d400,
1253 0x1d455, 0x1d456,
1254 0x1d49d, 0x1d49e,
1255 0x1d4a0, 0x1d4a2,
1256 0x1d4a3, 0x1d4a5,
1257 0x1d4a7, 0x1d4a9,
1258 0x1d4ad, 0x1d4ae,
1259 0x1d4ba, 0x1d4bb,
1260 0x1d4bc, 0x1d4bd,
1261 0x1d4c4, 0x1d4c5,
1262 0x1d506, 0x1d507,
1263 0x1d50b, 0x1d50d,
1264 0x1d515, 0x1d516,
1265 0x1d51d, 0x1d51e,
1266 0x1d53a, 0x1d53b,
1267 0x1d53f, 0x1d540,
1268 0x1d545, 0x1d546,
1269 0x1d547, 0x1d54a,
1270 0x1d551, 0x1d552,
1271 0x1d6a6, 0x1d6a8,
1272 0x1d6db, 0x1d6dc,
1273 0x1d715, 0x1d716,
1274 0x1d74f, 0x1d750,
1275 0x1d789, 0x1d78a,
1276 0x1d7c3, 0x1d7c4,
1277 0x1d7cc, 0x1d800,
1278 0x1da00, 0x1da37,
1279 0x1da3b, 0x1da6d,
1280 0x1da75, 0x1da76,
1281 0x1da84, 0x1da85,
1282 0x1da8c, 0x1e800,
1283 0x1e8d0, 0x1e8d7,
1284 0x1eef0, 0x1eef2,
1285 0x1f000, 0x1f110,
1286 0x1f12f, 0x1f130,
1287 0x1f16a, 0x1f170,
1288 0x1f19b, 0x1f1e6,
1289 0x1f203, 0x1f210,
1290 0x1f23b, 0x1f240,
1291 0x1f249, 0x1f250,
1292 0x1f252, 0x20000,
1293 0x2a6d7, 0x2a700,
1294 0x2b735, 0x2b740,
1295 0x2b81e, 0x2b820,
1296 0x2cea2, 0x2f800,
1297 0x2fa1e, 0xf0000,
1298 0xffffe, 0x100000,
1299 0x10fffe, 0x10ffff // sentinel
1300 };
1301
1302
1303 // use a binary search with a cache
1304
1305 private transient volatile int stCache = 0;
1306
1307 private boolean isStrongDirectional(char c) {
1308 int cachedIndex = stCache;
1309 if (c < strongTable[cachedIndex]) {
1310 cachedIndex = search(c, strongTable, 0, cachedIndex);
1311 } else if (c >= strongTable[cachedIndex + 1]) {
1312 cachedIndex = search(c, strongTable, cachedIndex + 1,
1313 strongTable.length - cachedIndex - 1);
1314 }
1315 boolean val = (cachedIndex & 0x1) == 1;
1316 stCache = cachedIndex;
1317 return val;
1318 }
1319
1320 private static int getKeyFromMask(int mask) {
1321 int key = 0;
1322 while (key < NUM_KEYS && ((mask & (1<<key)) == 0)) {
1323 ++key;
1324 }
1325 if (key == NUM_KEYS || ((mask & ~(1<<key)) != 0)) {
1326 throw new IllegalArgumentException("invalid shaper: " + Integer.toHexString(mask));
1327 }
1328 return key;
1329 }
1330
1331 /**
1332 * Returns a shaper for the provided unicode range. All
1333 * Latin-1 (EUROPEAN) digits are converted
1334 * to the corresponding decimal unicode digits.
1335 * @param singleRange the specified Unicode range
1336 * @return a non-contextual numeric shaper
1337 * @throws IllegalArgumentException if the range is not a single range
1338 */
1339 public static NumericShaper getShaper(int singleRange) {
1340 int key = getKeyFromMask(singleRange);
1341 return new NumericShaper(key, singleRange);
1342 }
1343
1344 /**
1345 * Returns a shaper for the provided Unicode
1346 * range. All Latin-1 (EUROPEAN) digits are converted to the
1347 * corresponding decimal digits of the specified Unicode range.
1348 *
1349 * @param singleRange the Unicode range given by a {@link
1350 * NumericShaper.Range} constant.
1351 * @return a non-contextual {@code NumericShaper}.
1352 * @throws NullPointerException if {@code singleRange} is {@code null}
1353 * @since 1.7
1354 */
1355 public static NumericShaper getShaper(Range singleRange) {
1356 return new NumericShaper(singleRange, EnumSet.of(singleRange));
1357 }
1358
1359 /**
1360 * Returns a contextual shaper for the provided unicode range(s).
1361 * Latin-1 (EUROPEAN) digits are converted to the decimal digits
1362 * corresponding to the range of the preceding text, if the
1363 * range is one of the provided ranges. Multiple ranges are
1364 * represented by or-ing the values together, such as,
1365 * {@code NumericShaper.ARABIC | NumericShaper.THAI}. The
1366 * shaper assumes EUROPEAN as the starting context, that is, if
1367 * EUROPEAN digits are encountered before any strong directional
1368 * text in the string, the context is presumed to be EUROPEAN, and
1369 * so the digits will not shape.
1370 * @param ranges the specified Unicode ranges
1371 * @return a shaper for the specified ranges
1372 */
1373 public static NumericShaper getContextualShaper(int ranges) {
1374 ranges |= CONTEXTUAL_MASK;
1375 return new NumericShaper(EUROPEAN_KEY, ranges);
1376 }
1377
1378 /**
1379 * Returns a contextual shaper for the provided Unicode
1380 * range(s). The Latin-1 (EUROPEAN) digits are converted to the
1381 * decimal digits corresponding to the range of the preceding
1382 * text, if the range is one of the provided ranges.
1383 *
1384 * <p>The shaper assumes EUROPEAN as the starting context, that
1385 * is, if EUROPEAN digits are encountered before any strong
1386 * directional text in the string, the context is presumed to be
1387 * EUROPEAN, and so the digits will not shape.
1388 *
1389 * @param ranges the specified Unicode ranges
1390 * @return a contextual shaper for the specified ranges
1391 * @throws NullPointerException if {@code ranges} is {@code null}.
1392 * @since 1.7
1393 */
1394 public static NumericShaper getContextualShaper(Set<Range> ranges) {
1395 NumericShaper shaper = new NumericShaper(Range.EUROPEAN, ranges);
1396 shaper.mask = CONTEXTUAL_MASK;
1397 return shaper;
1398 }
1399
1400 /**
1401 * Returns a contextual shaper for the provided unicode range(s).
1402 * Latin-1 (EUROPEAN) digits will be converted to the decimal digits
1403 * corresponding to the range of the preceding text, if the
1404 * range is one of the provided ranges. Multiple ranges are
1405 * represented by or-ing the values together, for example,
1406 * {@code NumericShaper.ARABIC | NumericShaper.THAI}. The
1407 * shaper uses defaultContext as the starting context.
1408 * @param ranges the specified Unicode ranges
1409 * @param defaultContext the starting context, such as
1410 * {@code NumericShaper.EUROPEAN}
1411 * @return a shaper for the specified Unicode ranges.
1412 * @throws IllegalArgumentException if the specified
1413 * {@code defaultContext} is not a single valid range.
1414 */
1415 public static NumericShaper getContextualShaper(int ranges, int defaultContext) {
1416 int key = getKeyFromMask(defaultContext);
1417 ranges |= CONTEXTUAL_MASK;
1418 return new NumericShaper(key, ranges);
1419 }
1420
1421 /**
1422 * Returns a contextual shaper for the provided Unicode range(s).
1423 * The Latin-1 (EUROPEAN) digits will be converted to the decimal
1424 * digits corresponding to the range of the preceding text, if the
1425 * range is one of the provided ranges. The shaper uses {@code
1426 * defaultContext} as the starting context.
1427 *
1428 * @param ranges the specified Unicode ranges
1429 * @param defaultContext the starting context, such as
1430 * {@code NumericShaper.Range.EUROPEAN}
1431 * @return a contextual shaper for the specified Unicode ranges.
1432 * @throws NullPointerException
1433 * if {@code ranges} or {@code defaultContext} is {@code null}
1434 * @since 1.7
1435 */
1436 public static NumericShaper getContextualShaper(Set<Range> ranges,
1437 Range defaultContext) {
1438 if (defaultContext == null) {
1439 throw new NullPointerException();
1440 }
1441 NumericShaper shaper = new NumericShaper(defaultContext, ranges);
1442 shaper.mask = CONTEXTUAL_MASK;
1443 return shaper;
1444 }
1445
1446 /**
1447 * Private constructor.
1448 */
1449 private NumericShaper(int key, int mask) {
1450 this.key = key;
1451 this.mask = mask;
1452 }
1453
1454 private NumericShaper(Range defaultContext, Set<Range> ranges) {
1455 shapingRange = defaultContext;
1456 rangeSet = EnumSet.copyOf(ranges); // throws NPE if ranges is null.
1457
1458 // Give precedence to EASTERN_ARABIC if both ARABIC and
1459 // EASTERN_ARABIC are specified.
1460 if (rangeSet.contains(Range.EASTERN_ARABIC)
1461 && rangeSet.contains(Range.ARABIC)) {
1462 rangeSet.remove(Range.ARABIC);
1463 }
1464
1465 // As well as the above case, give precedence to TAI_THAM_THAM if both
1466 // TAI_THAM_HORA and TAI_THAM_THAM are specified.
1467 if (rangeSet.contains(Range.TAI_THAM_THAM)
1468 && rangeSet.contains(Range.TAI_THAM_HORA)) {
1469 rangeSet.remove(Range.TAI_THAM_HORA);
1470 }
1471
1472 rangeArray = rangeSet.toArray(new Range[rangeSet.size()]);
1473 if (rangeArray.length > BSEARCH_THRESHOLD) {
1474 // sort rangeArray for binary search
1475 Arrays.sort(rangeArray,
1476 new Comparator<Range>() {
1477 public int compare(Range s1, Range s2) {
1478 return s1.base > s2.base ? 1 : s1.base == s2.base ? 0 : -1;
1479 }
1480 });
1481 }
1482 }
1483
1484 /**
1485 * Converts the digits in the text that occur between start and
1486 * start + count.
1487 * @param text an array of characters to convert
1488 * @param start the index into {@code text} to start
1489 * converting
1490 * @param count the number of characters in {@code text}
1491 * to convert
1492 * @throws IndexOutOfBoundsException if start or start + count is
1493 * out of bounds
1494 * @throws NullPointerException if text is null
1495 */
1496 public void shape(char[] text, int start, int count) {
1497 checkParams(text, start, count);
1498 if (isContextual()) {
1499 if (rangeSet == null) {
1500 shapeContextually(text, start, count, key);
1501 } else {
1502 shapeContextually(text, start, count, shapingRange);
1503 }
1504 } else {
1505 shapeNonContextually(text, start, count);
1506 }
1507 }
1508
1509 /**
1510 * Converts the digits in the text that occur between start and
1511 * start + count, using the provided context.
1512 * Context is ignored if the shaper is not a contextual shaper.
1513 * @param text an array of characters
1514 * @param start the index into {@code text} to start
1515 * converting
1516 * @param count the number of characters in {@code text}
1517 * to convert
1518 * @param context the context to which to convert the
1519 * characters, such as {@code NumericShaper.EUROPEAN}
1520 * @throws IndexOutOfBoundsException if start or start + count is
1521 * out of bounds
1522 * @throws NullPointerException if text is null
1523 * @throws IllegalArgumentException if this is a contextual shaper
1524 * and the specified {@code context} is not a single valid
1525 * range.
1526 */
1527 public void shape(char[] text, int start, int count, int context) {
1528 checkParams(text, start, count);
1529 if (isContextual()) {
1530 int ctxKey = getKeyFromMask(context);
1531 if (rangeSet == null) {
1532 shapeContextually(text, start, count, ctxKey);
1533 } else {
1534 shapeContextually(text, start, count, Range.values()[ctxKey]);
1535 }
1536 } else {
1537 shapeNonContextually(text, start, count);
1538 }
1539 }
1540
1541 /**
1542 * Converts the digits in the text that occur between {@code
1543 * start} and {@code start + count}, using the provided {@code
1544 * context}. {@code Context} is ignored if the shaper is not a
1545 * contextual shaper.
1546 *
1547 * @param text a {@code char} array
1548 * @param start the index into {@code text} to start converting
1549 * @param count the number of {@code char}s in {@code text}
1550 * to convert
1551 * @param context the context to which to convert the characters,
1552 * such as {@code NumericShaper.Range.EUROPEAN}
1553 * @throws IndexOutOfBoundsException
1554 * if {@code start} or {@code start + count} is out of bounds
1555 * @throws NullPointerException
1556 * if {@code text} or {@code context} is null
1557 * @since 1.7
1558 */
1559 public void shape(char[] text, int start, int count, Range context) {
1560 checkParams(text, start, count);
1561 if (context == null) {
1562 throw new NullPointerException("context is null");
1563 }
1564
1565 if (isContextual()) {
1566 if (rangeSet != null) {
1567 shapeContextually(text, start, count, context);
1568 } else {
1569 int key = Range.toRangeIndex(context);
1570 if (key >= 0) {
1571 shapeContextually(text, start, count, key);
1572 } else {
1573 shapeContextually(text, start, count, shapingRange);
1574 }
1575 }
1576 } else {
1577 shapeNonContextually(text, start, count);
1578 }
1579 }
1580
1581 private void checkParams(char[] text, int start, int count) {
1582 if (text == null) {
1583 throw new NullPointerException("text is null");
1584 }
1585 if ((start < 0)
1586 || (start > text.length)
1587 || ((start + count) < 0)
1588 || ((start + count) > text.length)) {
1589 throw new IndexOutOfBoundsException(
1590 "bad start or count for text of length " + text.length);
1591 }
1592 }
1593
1594 /**
1595 * Returns a {@code boolean} indicating whether or not
1596 * this shaper shapes contextually.
1597 * @return {@code true} if this shaper is contextual;
1598 * {@code false} otherwise.
1599 */
1600 public boolean isContextual() {
1601 return (mask & CONTEXTUAL_MASK) != 0;
1602 }
1603
1604 /**
1605 * Returns an {@code int} that ORs together the values for
1606 * all the ranges that will be shaped.
1607 * <p>
1608 * For example, to check if a shaper shapes to Arabic, you would use the
1609 * following:
1610 * <blockquote>
1611 * {@code if ((shaper.getRanges() & shaper.ARABIC) != 0) { ... }
1612 * </blockquote>
1613 *
1614 * <p>Note that this method supports only the bit mask-based
1615 * ranges. Call {@link #getRangeSet()} for the enum-based ranges.
1616 *
1617 * @return the values for all the ranges to be shaped.
1618 */
1619 public int getRanges() {
1620 return mask & ~CONTEXTUAL_MASK;
1621 }
1622
1623 /**
1624 * Returns a {@code Set} representing all the Unicode ranges in
1625 * this {@code NumericShaper} that will be shaped.
1626 *
1627 * @return all the Unicode ranges to be shaped.
1628 * @since 1.7
1629 */
1630 public Set<Range> getRangeSet() {
1631 if (rangeSet != null) {
1632 return EnumSet.copyOf(rangeSet);
1633 }
1634 return Range.maskToRangeSet(mask);
1635 }
1636
1637 /**
1638 * Perform non-contextual shaping.
1639 */
1640 private void shapeNonContextually(char[] text, int start, int count) {
1641 int base;
1642 char minDigit = '0';
1643 if (shapingRange != null) {
1644 base = shapingRange.getDigitBase();
1645 minDigit += shapingRange.getNumericBase();
1646 } else {
1647 base = bases[key];
1648 if (key == ETHIOPIC_KEY) {
1649 minDigit++; // Ethiopic doesn't use decimal zero
1650 }
1651 }
1652 for (int i = start, e = start + count; i < e; ++i) {
1653 char c = text[i];
1654 if (c >= minDigit && c <= '\u0039') {
1655 text[i] = (char)(c + base);
1656 }
1657 }
1658 }
1659
1660 /**
1661 * Perform contextual shaping.
1662 * Synchronized to protect caches used in getContextKey.
1663 */
1664 private synchronized void shapeContextually(char[] text, int start, int count, int ctxKey) {
1665
1666 // if we don't support this context, then don't shape
1667 if ((mask & (1<<ctxKey)) == 0) {
1668 ctxKey = EUROPEAN_KEY;
1669 }
1670 int lastkey = ctxKey;
1671
1672 int base = bases[ctxKey];
1673 char minDigit = ctxKey == ETHIOPIC_KEY ? '1' : '0'; // Ethiopic doesn't use decimal zero
1674
1675 synchronized (NumericShaper.class) {
1676 for (int i = start, e = start + count; i < e; ++i) {
1677 char c = text[i];
1678 if (c >= minDigit && c <= '\u0039') {
1679 text[i] = (char)(c + base);
1680 }
1681
1682 if (isStrongDirectional(c)) {
1683 int newkey = getContextKey(c);
1684 if (newkey != lastkey) {
1685 lastkey = newkey;
1686
1687 ctxKey = newkey;
1688 if (((mask & EASTERN_ARABIC) != 0) &&
1689 (ctxKey == ARABIC_KEY ||
1690 ctxKey == EASTERN_ARABIC_KEY)) {
1691 ctxKey = EASTERN_ARABIC_KEY;
1692 } else if (((mask & ARABIC) != 0) &&
1693 (ctxKey == ARABIC_KEY ||
1694 ctxKey == EASTERN_ARABIC_KEY)) {
1695 ctxKey = ARABIC_KEY;
1696 } else if ((mask & (1<<ctxKey)) == 0) {
1697 ctxKey = EUROPEAN_KEY;
1698 }
1699
1700 base = bases[ctxKey];
1701
1702 minDigit = ctxKey == ETHIOPIC_KEY ? '1' : '0'; // Ethiopic doesn't use decimal zero
1703 }
1704 }
1705 }
1706 }
1707 }
1708
1709 private void shapeContextually(char[] text, int start, int count, Range ctxKey) {
1710 // if we don't support the specified context, then don't shape.
1711 if (ctxKey == null || !rangeSet.contains(ctxKey)) {
1712 ctxKey = Range.EUROPEAN;
1713 }
1714
1715 Range lastKey = ctxKey;
1716 int base = ctxKey.getDigitBase();
1717 char minDigit = (char)('0' + ctxKey.getNumericBase());
1718 final int end = start + count;
1719 for (int i = start; i < end; ++i) {
1720 char c = text[i];
1721 if (c >= minDigit && c <= '9') {
1722 text[i] = (char)(c + base);
1723 continue;
1724 }
1725 if (isStrongDirectional(c)) {
1726 ctxKey = rangeForCodePoint(c);
1727 if (ctxKey != lastKey) {
1728 lastKey = ctxKey;
1729 base = ctxKey.getDigitBase();
1730 minDigit = (char)('0' + ctxKey.getNumericBase());
1731 }
1732 }
1733 }
1734 }
1735
1736 /**
1737 * Returns a hash code for this shaper.
1738 * @return this shaper's hash code.
1739 * @see java.lang.Object#hashCode
1740 */
1741 public int hashCode() {
1742 int hash = mask;
1743 if (rangeSet != null) {
1744 // Use the CONTEXTUAL_MASK bit only for the enum-based
1745 // NumericShaper. A deserialized NumericShaper might have
1746 // bit masks.
1747 hash &= CONTEXTUAL_MASK;
1748 hash ^= rangeSet.hashCode();
1749 }
1750 return hash;
1751 }
1752
1753 /**
1754 * Returns {@code true} if the specified object is an instance of
1755 * {@code NumericShaper} and shapes identically to this one,
1756 * regardless of the range representations, the bit mask or the
1757 * enum. For example, the following code produces {@code "true"}.
1758 * <blockquote><pre>
1759 * NumericShaper ns1 = NumericShaper.getShaper(NumericShaper.ARABIC);
1760 * NumericShaper ns2 = NumericShaper.getShaper(NumericShaper.Range.ARABIC);
1761 * System.out.println(ns1.equals(ns2));
1762 * </pre></blockquote>
1763 *
1764 * @param o the specified object to compare to this
1765 * {@code NumericShaper}
1766 * @return {@code true} if {@code o} is an instance
1767 * of {@code NumericShaper} and shapes in the same way;
1768 * {@code false} otherwise.
1769 * @see java.lang.Object#equals(java.lang.Object)
1770 */
1771 public boolean equals(Object o) {
1772 if (o != null) {
1773 try {
1774 NumericShaper rhs = (NumericShaper)o;
1775 if (rangeSet != null) {
1776 if (rhs.rangeSet != null) {
1777 return isContextual() == rhs.isContextual()
1778 && rangeSet.equals(rhs.rangeSet)
1779 && shapingRange == rhs.shapingRange;
1780 }
1781 return isContextual() == rhs.isContextual()
1782 && rangeSet.equals(Range.maskToRangeSet(rhs.mask))
1783 && shapingRange == Range.indexToRange(rhs.key);
1784 } else if (rhs.rangeSet != null) {
1785 Set<Range> rset = Range.maskToRangeSet(mask);
1786 Range srange = Range.indexToRange(key);
1787 return isContextual() == rhs.isContextual()
1788 && rset.equals(rhs.rangeSet)
1789 && srange == rhs.shapingRange;
1790 }
1791 return rhs.mask == mask && rhs.key == key;
1792 }
1793 catch (ClassCastException e) {
1794 }
1795 }
1796 return false;
1797 }
1798
1799 /**
1800 * Returns a {@code String} that describes this shaper. This method
1801 * is used for debugging purposes only.
1802 * @return a {@code String} describing this shaper.
1803 */
1804 public String toString() {
1805 StringBuilder buf = new StringBuilder(super.toString());
1806
1807 buf.append("[contextual:").append(isContextual());
1808
1809 String[] keyNames = null;
1810 if (isContextual()) {
1811 buf.append(", context:");
1812 buf.append(shapingRange == null ? Range.values()[key] : shapingRange);
1813 }
1814
1815 if (rangeSet == null) {
1816 buf.append(", range(s): ");
1817 boolean first = true;
1818 for (int i = 0; i < NUM_KEYS; ++i) {
1819 if ((mask & (1 << i)) != 0) {
1820 if (first) {
1821 first = false;
1822 } else {
1823 buf.append(", ");
1824 }
1825 buf.append(Range.values()[i]);
1826 }
1827 }
1828 } else {
1829 buf.append(", range set: ").append(rangeSet);
1830 }
1831 buf.append(']');
1832
1833 return buf.toString();
1834 }
1835
1836 /**
1837 * Returns the index of the high bit in value (assuming le, actually
1838 * power of 2 >= value). value must be positive.
1839 */
1840 private static int getHighBit(int value) {
1841 if (value <= 0) {
1842 return -32;
1843 }
1844
1845 int bit = 0;
1846
1847 if (value >= 1 << 16) {
1848 value >>= 16;
1849 bit += 16;
1850 }
1851
1852 if (value >= 1 << 8) {
1853 value >>= 8;
1854 bit += 8;
1855 }
1856
1857 if (value >= 1 << 4) {
1858 value >>= 4;
1859 bit += 4;
1860 }
1861
1862 if (value >= 1 << 2) {
1863 value >>= 2;
1864 bit += 2;
1865 }
1866
1867 if (value >= 1 << 1) {
1868 bit += 1;
1869 }
1870
1871 return bit;
1872 }
1873
1874 /**
1875 * fast binary search over subrange of array.
1876 */
1877 private static int search(int value, int[] array, int start, int length)
1878 {
1879 int power = 1 << getHighBit(length);
1880 int extra = length - power;
1881 int probe = power;
1882 int index = start;
1883
1884 if (value >= array[index + extra]) {
1885 index += extra;
1886 }
1887
1888 while (probe > 1) {
1889 probe >>= 1;
1890
1891 if (value >= array[index + probe]) {
1892 index += probe;
1893 }
1894 }
1895
1896 return index;
1897 }
1898
1899 /**
1900 * Converts the {@code NumericShaper.Range} enum-based parameters,
1901 * if any, to the bit mask-based counterparts and writes this
1902 * object to the {@code stream}. Any enum constants that have no
1903 * bit mask-based counterparts are ignored in the conversion.
1904 *
1905 * @param stream the output stream to write to
1906 * @throws IOException if an I/O error occurs while writing to {@code stream}
1907 * @since 1.7
1908 */
1909 private void writeObject(ObjectOutputStream stream) throws IOException {
1910 if (shapingRange != null) {
1911 int index = Range.toRangeIndex(shapingRange);
1912 if (index >= 0) {
1913 key = index;
1914 }
1915 }
1916 if (rangeSet != null) {
1917 mask |= Range.toRangeMask(rangeSet);
1918 }
1919 stream.defaultWriteObject();
1920 }
1921 }