1 /* 2 * Copyright (c) 2005, 2011, 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 ******************************************************************************* 27 * (C) Copyright IBM Corp. and others, 1996-2009 - All Rights Reserved * 28 * * 29 * The original version of this source code and documentation is copyrighted * 30 * and owned by IBM, These materials are provided under terms of a License * 31 * Agreement between IBM and Sun. This technology is protected by multiple * 32 * US and International patents. This notice and attribution to IBM may not * 33 * to removed. * 34 ******************************************************************************* 35 */ 36 37 package sun.text.normalizer; 38 39 import java.text.CharacterIterator; 40 import java.text.Normalizer; 41 42 /** 43 * Unicode Normalization 44 * 45 * <h2>Unicode normalization API</h2> 46 * 47 * <code>normalize</code> transforms Unicode text into an equivalent composed or 48 * decomposed form, allowing for easier sorting and searching of text. 49 * <code>normalize</code> supports the standard normalization forms described in 50 * <a href="http://www.unicode.org/unicode/reports/tr15/" target="unicode"> 51 * Unicode Standard Annex #15 — Unicode Normalization Forms</a>. 52 * 53 * Characters with accents or other adornments can be encoded in 54 * several different ways in Unicode. For example, take the character A-acute. 55 * In Unicode, this can be encoded as a single character (the 56 * "composed" form): 57 * 58 * <p> 59 * 00C1 LATIN CAPITAL LETTER A WITH ACUTE 60 * </p> 61 * 62 * or as two separate characters (the "decomposed" form): 63 * 64 * <p> 65 * 0041 LATIN CAPITAL LETTER A 66 * 0301 COMBINING ACUTE ACCENT 67 * </p> 68 * 69 * To a user of your program, however, both of these sequences should be 70 * treated as the same "user-level" character "A with acute accent". When you 71 * are searching or comparing text, you must ensure that these two sequences are 72 * treated equivalently. In addition, you must handle characters with more than 73 * one accent. Sometimes the order of a character's combining accents is 74 * significant, while in other cases accent sequences in different orders are 75 * really equivalent. 76 * 77 * Similarly, the string "ffi" can be encoded as three separate letters: 78 * 79 * <p> 80 * 0066 LATIN SMALL LETTER F 81 * 0066 LATIN SMALL LETTER F 82 * 0069 LATIN SMALL LETTER I 83 * </p> 84 * 85 * or as the single character 86 * 87 * <p> 88 * FB03 LATIN SMALL LIGATURE FFI 89 * </p> 90 * 91 * The ffi ligature is not a distinct semantic character, and strictly speaking 92 * it shouldn't be in Unicode at all, but it was included for compatibility 93 * with existing character sets that already provided it. The Unicode standard 94 * identifies such characters by giving them "compatibility" decompositions 95 * into the corresponding semantic characters. When sorting and searching, you 96 * will often want to use these mappings. 97 * 98 * <code>normalize</code> helps solve these problems by transforming text into 99 * the canonical composed and decomposed forms as shown in the first example 100 * above. In addition, you can have it perform compatibility decompositions so 101 * that you can treat compatibility characters the same as their equivalents. 102 * Finally, <code>normalize</code> rearranges accents into the proper canonical 103 * order, so that you do not have to worry about accent rearrangement on your 104 * own. 105 * 106 * Form FCD, "Fast C or D", is also designed for collation. 107 * It allows to work on strings that are not necessarily normalized 108 * with an algorithm (like in collation) that works under "canonical closure", 109 * i.e., it treats precomposed characters and their decomposed equivalents the 110 * same. 111 * 112 * It is not a normalization form because it does not provide for uniqueness of 113 * representation. Multiple strings may be canonically equivalent (their NFDs 114 * are identical) and may all conform to FCD without being identical themselves. 115 * 116 * The form is defined such that the "raw decomposition", the recursive 117 * canonical decomposition of each character, results in a string that is 118 * canonically ordered. This means that precomposed characters are allowed for 119 * as long as their decompositions do not need canonical reordering. 120 * 121 * Its advantage for a process like collation is that all NFD and most NFC texts 122 * - and many unnormalized texts - already conform to FCD and do not need to be 123 * normalized (NFD) for such a process. The FCD quick check will return YES for 124 * most strings in practice. 125 * 126 * normalize(FCD) may be implemented with NFD. 127 * 128 * For more details on FCD see the collation design document: 129 * http://source.icu-project.org/repos/icu/icuhtml/trunk/design/collation/ICU_collation_design.htm 130 * 131 * ICU collation performs either NFD or FCD normalization automatically if 132 * normalization is turned on for the collator object. Beyond collation and 133 * string search, normalized strings may be useful for string equivalence 134 * comparisons, transliteration/transcription, unique representations, etc. 135 * 136 * The W3C generally recommends to exchange texts in NFC. 137 * Note also that most legacy character encodings use only precomposed forms and 138 * often do not encode any combining marks by themselves. For conversion to such 139 * character encodings the Unicode text needs to be normalized to NFC. 140 * For more usage examples, see the Unicode Standard Annex. 141 * @stable ICU 2.8 142 */ 143 144 public final class NormalizerBase implements Cloneable { 145 146 //------------------------------------------------------------------------- 147 // Private data 148 //------------------------------------------------------------------------- 149 private char[] buffer = new char[100]; 150 private int bufferStart = 0; 151 private int bufferPos = 0; 152 private int bufferLimit = 0; 153 154 // The input text and our position in it 155 private UCharacterIterator text; 156 private Mode mode = NFC; 157 private int options = 0; 158 private int currentIndex; 159 private int nextIndex; 160 161 /** 162 * Options bit set value to select Unicode 3.2 normalization 163 * (except NormalizationCorrections). 164 * At most one Unicode version can be selected at a time. 165 * @stable ICU 2.6 166 */ 167 public static final int UNICODE_3_2=0x20; 168 169 /** 170 * Constant indicating that the end of the iteration has been reached. 171 * This is guaranteed to have the same value as {@link UCharacterIterator#DONE}. 172 * @stable ICU 2.8 173 */ 174 public static final int DONE = UCharacterIterator.DONE; 175 176 /** 177 * Constants for normalization modes. 178 * @stable ICU 2.8 179 */ 180 public static class Mode { 181 private int modeValue; 182 private Mode(int value) { 183 modeValue = value; 184 } 185 186 /** 187 * This method is used for method dispatch 188 * @stable ICU 2.6 189 */ 190 protected int normalize(char[] src, int srcStart, int srcLimit, 191 char[] dest,int destStart,int destLimit, 192 UnicodeSet nx) { 193 int srcLen = (srcLimit - srcStart); 194 int destLen = (destLimit - destStart); 195 if( srcLen > destLen ) { 196 return srcLen; 197 } 198 System.arraycopy(src,srcStart,dest,destStart,srcLen); 199 return srcLen; 200 } 201 202 /** 203 * This method is used for method dispatch 204 * @stable ICU 2.6 205 */ 206 protected int normalize(char[] src, int srcStart, int srcLimit, 207 char[] dest,int destStart,int destLimit, 208 int options) { 209 return normalize( src, srcStart, srcLimit, 210 dest,destStart,destLimit, 211 NormalizerImpl.getNX(options) 212 ); 213 } 214 215 /** 216 * This method is used for method dispatch 217 * @stable ICU 2.6 218 */ 219 protected String normalize(String src, int options) { 220 return src; 221 } 222 223 /** 224 * This method is used for method dispatch 225 * @stable ICU 2.8 226 */ 227 protected int getMinC() { 228 return -1; 229 } 230 231 /** 232 * This method is used for method dispatch 233 * @stable ICU 2.8 234 */ 235 protected int getMask() { 236 return -1; 237 } 238 239 /** 240 * This method is used for method dispatch 241 * @stable ICU 2.8 242 */ 243 protected IsPrevBoundary getPrevBoundary() { 244 return null; 245 } 246 247 /** 248 * This method is used for method dispatch 249 * @stable ICU 2.8 250 */ 251 protected IsNextBoundary getNextBoundary() { 252 return null; 253 } 254 255 /** 256 * This method is used for method dispatch 257 * @stable ICU 2.6 258 */ 259 protected QuickCheckResult quickCheck(char[] src,int start, int limit, 260 boolean allowMaybe,UnicodeSet nx) { 261 if(allowMaybe) { 262 return MAYBE; 263 } 264 return NO; 265 } 266 267 /** 268 * This method is used for method dispatch 269 * @stable ICU 2.8 270 */ 271 protected boolean isNFSkippable(int c) { 272 return true; 273 } 274 } 275 276 /** 277 * No decomposition/composition. 278 * @stable ICU 2.8 279 */ 280 public static final Mode NONE = new Mode(1); 281 282 /** 283 * Canonical decomposition. 284 * @stable ICU 2.8 285 */ 286 public static final Mode NFD = new NFDMode(2); 287 288 private static final class NFDMode extends Mode { 289 private NFDMode(int value) { 290 super(value); 291 } 292 293 protected int normalize(char[] src, int srcStart, int srcLimit, 294 char[] dest,int destStart,int destLimit, 295 UnicodeSet nx) { 296 int[] trailCC = new int[1]; 297 return NormalizerImpl.decompose(src, srcStart,srcLimit, 298 dest, destStart,destLimit, 299 false, trailCC,nx); 300 } 301 302 protected String normalize( String src, int options) { 303 return decompose(src,false,options); 304 } 305 306 protected int getMinC() { 307 return NormalizerImpl.MIN_WITH_LEAD_CC; 308 } 309 310 protected IsPrevBoundary getPrevBoundary() { 311 return new IsPrevNFDSafe(); 312 } 313 314 protected IsNextBoundary getNextBoundary() { 315 return new IsNextNFDSafe(); 316 } 317 318 protected int getMask() { 319 return (NormalizerImpl.CC_MASK|NormalizerImpl.QC_NFD); 320 } 321 322 protected QuickCheckResult quickCheck(char[] src,int start, 323 int limit,boolean allowMaybe, 324 UnicodeSet nx) { 325 return NormalizerImpl.quickCheck( 326 src, start,limit, 327 NormalizerImpl.getFromIndexesArr( 328 NormalizerImpl.INDEX_MIN_NFD_NO_MAYBE 329 ), 330 NormalizerImpl.QC_NFD, 331 0, 332 allowMaybe, 333 nx 334 ); 335 } 336 337 protected boolean isNFSkippable(int c) { 338 return NormalizerImpl.isNFSkippable(c,this, 339 (NormalizerImpl.CC_MASK|NormalizerImpl.QC_NFD) 340 ); 341 } 342 } 343 344 /** 345 * Compatibility decomposition. 346 * @stable ICU 2.8 347 */ 348 public static final Mode NFKD = new NFKDMode(3); 349 350 private static final class NFKDMode extends Mode { 351 private NFKDMode(int value) { 352 super(value); 353 } 354 355 protected int normalize(char[] src, int srcStart, int srcLimit, 356 char[] dest,int destStart,int destLimit, 357 UnicodeSet nx) { 358 int[] trailCC = new int[1]; 359 return NormalizerImpl.decompose(src, srcStart,srcLimit, 360 dest, destStart,destLimit, 361 true, trailCC, nx); 362 } 363 364 protected String normalize( String src, int options) { 365 return decompose(src,true,options); 366 } 367 368 protected int getMinC() { 369 return NormalizerImpl.MIN_WITH_LEAD_CC; 370 } 371 372 protected IsPrevBoundary getPrevBoundary() { 373 return new IsPrevNFDSafe(); 374 } 375 376 protected IsNextBoundary getNextBoundary() { 377 return new IsNextNFDSafe(); 378 } 379 380 protected int getMask() { 381 return (NormalizerImpl.CC_MASK|NormalizerImpl.QC_NFKD); 382 } 383 384 protected QuickCheckResult quickCheck(char[] src,int start, 385 int limit,boolean allowMaybe, 386 UnicodeSet nx) { 387 return NormalizerImpl.quickCheck( 388 src,start,limit, 389 NormalizerImpl.getFromIndexesArr( 390 NormalizerImpl.INDEX_MIN_NFKD_NO_MAYBE 391 ), 392 NormalizerImpl.QC_NFKD, 393 NormalizerImpl.OPTIONS_COMPAT, 394 allowMaybe, 395 nx 396 ); 397 } 398 399 protected boolean isNFSkippable(int c) { 400 return NormalizerImpl.isNFSkippable(c, this, 401 (NormalizerImpl.CC_MASK|NormalizerImpl.QC_NFKD) 402 ); 403 } 404 } 405 406 /** 407 * Canonical decomposition followed by canonical composition. 408 * @stable ICU 2.8 409 */ 410 public static final Mode NFC = new NFCMode(4); 411 412 private static final class NFCMode extends Mode{ 413 private NFCMode(int value) { 414 super(value); 415 } 416 protected int normalize(char[] src, int srcStart, int srcLimit, 417 char[] dest,int destStart,int destLimit, 418 UnicodeSet nx) { 419 return NormalizerImpl.compose( src, srcStart, srcLimit, 420 dest,destStart,destLimit, 421 0, nx); 422 } 423 424 protected String normalize( String src, int options) { 425 return compose(src, false, options); 426 } 427 428 protected int getMinC() { 429 return NormalizerImpl.getFromIndexesArr( 430 NormalizerImpl.INDEX_MIN_NFC_NO_MAYBE 431 ); 432 } 433 protected IsPrevBoundary getPrevBoundary() { 434 return new IsPrevTrueStarter(); 435 } 436 protected IsNextBoundary getNextBoundary() { 437 return new IsNextTrueStarter(); 438 } 439 protected int getMask() { 440 return (NormalizerImpl.CC_MASK|NormalizerImpl.QC_NFC); 441 } 442 protected QuickCheckResult quickCheck(char[] src,int start, 443 int limit,boolean allowMaybe, 444 UnicodeSet nx) { 445 return NormalizerImpl.quickCheck( 446 src,start,limit, 447 NormalizerImpl.getFromIndexesArr( 448 NormalizerImpl.INDEX_MIN_NFC_NO_MAYBE 449 ), 450 NormalizerImpl.QC_NFC, 451 0, 452 allowMaybe, 453 nx 454 ); 455 } 456 protected boolean isNFSkippable(int c) { 457 return NormalizerImpl.isNFSkippable(c,this, 458 ( NormalizerImpl.CC_MASK|NormalizerImpl.COMBINES_ANY| 459 (NormalizerImpl.QC_NFC & NormalizerImpl.QC_ANY_NO) 460 ) 461 ); 462 } 463 }; 464 465 /** 466 * Compatibility decomposition followed by canonical composition. 467 * @stable ICU 2.8 468 */ 469 public static final Mode NFKC =new NFKCMode(5); 470 471 private static final class NFKCMode extends Mode{ 472 private NFKCMode(int value) { 473 super(value); 474 } 475 protected int normalize(char[] src, int srcStart, int srcLimit, 476 char[] dest,int destStart,int destLimit, 477 UnicodeSet nx) { 478 return NormalizerImpl.compose(src, srcStart,srcLimit, 479 dest, destStart,destLimit, 480 NormalizerImpl.OPTIONS_COMPAT, nx); 481 } 482 483 protected String normalize( String src, int options) { 484 return compose(src, true, options); 485 } 486 protected int getMinC() { 487 return NormalizerImpl.getFromIndexesArr( 488 NormalizerImpl.INDEX_MIN_NFKC_NO_MAYBE 489 ); 490 } 491 protected IsPrevBoundary getPrevBoundary() { 492 return new IsPrevTrueStarter(); 493 } 494 protected IsNextBoundary getNextBoundary() { 495 return new IsNextTrueStarter(); 496 } 497 protected int getMask() { 498 return (NormalizerImpl.CC_MASK|NormalizerImpl.QC_NFKC); 499 } 500 protected QuickCheckResult quickCheck(char[] src,int start, 501 int limit,boolean allowMaybe, 502 UnicodeSet nx) { 503 return NormalizerImpl.quickCheck( 504 src,start,limit, 505 NormalizerImpl.getFromIndexesArr( 506 NormalizerImpl.INDEX_MIN_NFKC_NO_MAYBE 507 ), 508 NormalizerImpl.QC_NFKC, 509 NormalizerImpl.OPTIONS_COMPAT, 510 allowMaybe, 511 nx 512 ); 513 } 514 protected boolean isNFSkippable(int c) { 515 return NormalizerImpl.isNFSkippable(c, this, 516 ( NormalizerImpl.CC_MASK|NormalizerImpl.COMBINES_ANY| 517 (NormalizerImpl.QC_NFKC & NormalizerImpl.QC_ANY_NO) 518 ) 519 ); 520 } 521 }; 522 523 /** 524 * Result values for quickCheck(). 525 * For details see Unicode Technical Report 15. 526 * @stable ICU 2.8 527 */ 528 public static final class QuickCheckResult{ 529 private int resultValue; 530 private QuickCheckResult(int value) { 531 resultValue=value; 532 } 533 } 534 /** 535 * Indicates that string is not in the normalized format 536 * @stable ICU 2.8 537 */ 538 public static final QuickCheckResult NO = new QuickCheckResult(0); 539 540 /** 541 * Indicates that string is in the normalized format 542 * @stable ICU 2.8 543 */ 544 public static final QuickCheckResult YES = new QuickCheckResult(1); 545 546 /** 547 * Indicates it cannot be determined if string is in the normalized 548 * format without further thorough checks. 549 * @stable ICU 2.8 550 */ 551 public static final QuickCheckResult MAYBE = new QuickCheckResult(2); 552 553 //------------------------------------------------------------------------- 554 // Constructors 555 //------------------------------------------------------------------------- 556 557 /** 558 * Creates a new <tt>Normalizer</tt> object for iterating over the 559 * normalized form of a given string. 560 * <p> 561 * The <tt>options</tt> parameter specifies which optional 562 * <tt>Normalizer</tt> features are to be enabled for this object. 563 * <p> 564 * @param str The string to be normalized. The normalization 565 * will start at the beginning of the string. 566 * 567 * @param mode The normalization mode. 568 * 569 * @param opt Any optional features to be enabled. 570 * Currently the only available option is {@link #UNICODE_3_2}. 571 * If you want the default behavior corresponding to one of the 572 * standard Unicode Normalization Forms, use 0 for this argument. 573 * @stable ICU 2.6 574 */ 575 public NormalizerBase(String str, Mode mode, int opt) { 576 this.text = UCharacterIterator.getInstance(str); 577 this.mode = mode; 578 this.options=opt; 579 } 580 581 /** 582 * Creates a new <tt>Normalizer</tt> object for iterating over the 583 * normalized form of the given text. 584 * <p> 585 * @param iter The input text to be normalized. The normalization 586 * will start at the beginning of the string. 587 * 588 * @param mode The normalization mode. 589 */ 590 public NormalizerBase(CharacterIterator iter, Mode mode) { 591 this(iter, mode, UNICODE_LATEST); 592 } 593 594 /** 595 * Creates a new <tt>Normalizer</tt> object for iterating over the 596 * normalized form of the given text. 597 * <p> 598 * @param iter The input text to be normalized. The normalization 599 * will start at the beginning of the string. 600 * 601 * @param mode The normalization mode. 602 * 603 * @param opt Any optional features to be enabled. 604 * Currently the only available option is {@link #UNICODE_3_2}. 605 * If you want the default behavior corresponding to one of the 606 * standard Unicode Normalization Forms, use 0 for this argument. 607 * @stable ICU 2.6 608 */ 609 public NormalizerBase(CharacterIterator iter, Mode mode, int opt) { 610 this.text = UCharacterIterator.getInstance( 611 (CharacterIterator)iter.clone() 612 ); 613 this.mode = mode; 614 this.options = opt; 615 } 616 617 /** 618 * Clones this <tt>Normalizer</tt> object. All properties of this 619 * object are duplicated in the new object, including the cloning of any 620 * {@link CharacterIterator} that was passed in to the constructor 621 * or to {@link #setText(CharacterIterator) setText}. 622 * However, the text storage underlying 623 * the <tt>CharacterIterator</tt> is not duplicated unless the 624 * iterator's <tt>clone</tt> method does so. 625 * @stable ICU 2.8 626 */ 627 public Object clone() { 628 try { 629 NormalizerBase copy = (NormalizerBase) super.clone(); 630 copy.text = (UCharacterIterator) text.clone(); 631 //clone the internal buffer 632 if (buffer != null) { 633 copy.buffer = new char[buffer.length]; 634 System.arraycopy(buffer,0,copy.buffer,0,buffer.length); 635 } 636 return copy; 637 } 638 catch (CloneNotSupportedException e) { 639 throw new InternalError(e.toString(), e); 640 } 641 } 642 643 //-------------------------------------------------------------------------- 644 // Static Utility methods 645 //-------------------------------------------------------------------------- 646 647 /** 648 * Compose a string. 649 * The string will be composed according to the specified mode. 650 * @param str The string to compose. 651 * @param compat If true the string will be composed according to 652 * NFKC rules and if false will be composed according to 653 * NFC rules. 654 * @param options The only recognized option is UNICODE_3_2 655 * @return String The composed string 656 * @stable ICU 2.6 657 */ 658 public static String compose(String str, boolean compat, int options) { 659 660 char[] dest, src; 661 if (options == UNICODE_3_2_0_ORIGINAL) { 662 String mappedStr = NormalizerImpl.convert(str); 663 dest = new char[mappedStr.length()*MAX_BUF_SIZE_COMPOSE]; 664 src = mappedStr.toCharArray(); 665 } else { 666 dest = new char[str.length()*MAX_BUF_SIZE_COMPOSE]; 667 src = str.toCharArray(); 668 } 669 int destSize=0; 670 671 UnicodeSet nx = NormalizerImpl.getNX(options); 672 673 /* reset options bits that should only be set here or inside compose() */ 674 options&=~(NormalizerImpl.OPTIONS_SETS_MASK|NormalizerImpl.OPTIONS_COMPAT|NormalizerImpl.OPTIONS_COMPOSE_CONTIGUOUS); 675 676 if(compat) { 677 options|=NormalizerImpl.OPTIONS_COMPAT; 678 } 679 680 for(;;) { 681 destSize=NormalizerImpl.compose(src,0,src.length, 682 dest,0,dest.length,options, 683 nx); 684 if(destSize<=dest.length) { 685 return new String(dest,0,destSize); 686 } else { 687 dest = new char[destSize]; 688 } 689 } 690 } 691 692 private static final int MAX_BUF_SIZE_COMPOSE = 2; 693 private static final int MAX_BUF_SIZE_DECOMPOSE = 3; 694 695 /** 696 * Decompose a string. 697 * The string will be decomposed according to the specified mode. 698 * @param str The string to decompose. 699 * @param compat If true the string will be decomposed according to NFKD 700 * rules and if false will be decomposed according to NFD 701 * rules. 702 * @return String The decomposed string 703 * @stable ICU 2.8 704 */ 705 public static String decompose(String str, boolean compat) { 706 return decompose(str,compat,UNICODE_LATEST); 707 } 708 709 /** 710 * Decompose a string. 711 * The string will be decomposed according to the specified mode. 712 * @param str The string to decompose. 713 * @param compat If true the string will be decomposed according to NFKD 714 * rules and if false will be decomposed according to NFD 715 * rules. 716 * @param options The normalization options, ORed together (0 for no options). 717 * @return String The decomposed string 718 * @stable ICU 2.6 719 */ 720 public static String decompose(String str, boolean compat, int options) { 721 722 int[] trailCC = new int[1]; 723 int destSize=0; 724 UnicodeSet nx = NormalizerImpl.getNX(options); 725 char[] dest; 726 727 if (options == UNICODE_3_2_0_ORIGINAL) { 728 String mappedStr = NormalizerImpl.convert(str); 729 dest = new char[mappedStr.length()*MAX_BUF_SIZE_DECOMPOSE]; 730 731 for(;;) { 732 destSize=NormalizerImpl.decompose(mappedStr.toCharArray(),0,mappedStr.length(), 733 dest,0,dest.length, 734 compat,trailCC, nx); 735 if(destSize<=dest.length) { 736 return new String(dest,0,destSize); 737 } else { 738 dest = new char[destSize]; 739 } 740 } 741 } else { 742 dest = new char[str.length()*MAX_BUF_SIZE_DECOMPOSE]; 743 744 for(;;) { 745 destSize=NormalizerImpl.decompose(str.toCharArray(),0,str.length(), 746 dest,0,dest.length, 747 compat,trailCC, nx); 748 if(destSize<=dest.length) { 749 return new String(dest,0,destSize); 750 } else { 751 dest = new char[destSize]; 752 } 753 } 754 } 755 } 756 757 /** 758 * Normalize a string. 759 * The string will be normalized according to the specified normalization 760 * mode and options. 761 * @param src The char array to compose. 762 * @param srcStart Start index of the source 763 * @param srcLimit Limit index of the source 764 * @param dest The char buffer to fill in 765 * @param destStart Start index of the destination buffer 766 * @param destLimit End index of the destination buffer 767 * @param mode The normalization mode; one of Normalizer.NONE, 768 * Normalizer.NFD, Normalizer.NFC, Normalizer.NFKC, 769 * Normalizer.NFKD, Normalizer.DEFAULT 770 * @param options The normalization options, ORed together (0 for no options). 771 * @return int The total buffer size needed;if greater than length of 772 * result, the output was truncated. 773 * @exception IndexOutOfBoundsException if the target capacity is 774 * less than the required length 775 * @stable ICU 2.6 776 */ 777 public static int normalize(char[] src,int srcStart, int srcLimit, 778 char[] dest,int destStart, int destLimit, 779 Mode mode, int options) { 780 int length = mode.normalize(src,srcStart,srcLimit,dest,destStart,destLimit, options); 781 782 if(length<=(destLimit-destStart)) { 783 return length; 784 } else { 785 throw new IndexOutOfBoundsException(Integer.toString(length)); 786 } 787 } 788 789 //------------------------------------------------------------------------- 790 // Iteration API 791 //------------------------------------------------------------------------- 792 793 /** 794 * Return the current character in the normalized text-> 795 * @return The codepoint as an int 796 * @stable ICU 2.8 797 */ 798 public int current() { 799 if(bufferPos<bufferLimit || nextNormalize()) { 800 return getCodePointAt(bufferPos); 801 } else { 802 return DONE; 803 } 804 } 805 806 /** 807 * Return the next character in the normalized text and advance 808 * the iteration position by one. If the end 809 * of the text has already been reached, {@link #DONE} is returned. 810 * @return The codepoint as an int 811 * @stable ICU 2.8 812 */ 813 public int next() { 814 if(bufferPos<bufferLimit || nextNormalize()) { 815 int c=getCodePointAt(bufferPos); 816 bufferPos+=(c>0xFFFF) ? 2 : 1; 817 return c; 818 } else { 819 return DONE; 820 } 821 } 822 823 824 /** 825 * Return the previous character in the normalized text and decrement 826 * the iteration position by one. If the beginning 827 * of the text has already been reached, {@link #DONE} is returned. 828 * @return The codepoint as an int 829 * @stable ICU 2.8 830 */ 831 public int previous() { 832 if(bufferPos>0 || previousNormalize()) { 833 int c=getCodePointAt(bufferPos-1); 834 bufferPos-=(c>0xFFFF) ? 2 : 1; 835 return c; 836 } else { 837 return DONE; 838 } 839 } 840 841 /** 842 * Reset the index to the beginning of the text. 843 * This is equivalent to setIndexOnly(startIndex)). 844 * @stable ICU 2.8 845 */ 846 public void reset() { 847 text.setIndex(0); 848 currentIndex=nextIndex=0; 849 clearBuffer(); 850 } 851 852 /** 853 * Set the iteration position in the input text that is being normalized, 854 * without any immediate normalization. 855 * After setIndexOnly(), getIndex() will return the same index that is 856 * specified here. 857 * 858 * @param index the desired index in the input text. 859 * @stable ICU 2.8 860 */ 861 public void setIndexOnly(int index) { 862 text.setIndex(index); 863 currentIndex=nextIndex=index; // validates index 864 clearBuffer(); 865 } 866 867 /** 868 * Set the iteration position in the input text that is being normalized 869 * and return the first normalized character at that position. 870 * <p> 871 * <b>Note:</b> This method sets the position in the <em>input</em> text, 872 * while {@link #next} and {@link #previous} iterate through characters 873 * in the normalized <em>output</em>. This means that there is not 874 * necessarily a one-to-one correspondence between characters returned 875 * by <tt>next</tt> and <tt>previous</tt> and the indices passed to and 876 * returned from <tt>setIndex</tt> and {@link #getIndex}. 877 * <p> 878 * @param index the desired index in the input text-> 879 * 880 * @return the first normalized character that is the result of iterating 881 * forward starting at the given index. 882 * 883 * @throws IllegalArgumentException if the given index is less than 884 * {@link #getBeginIndex} or greater than {@link #getEndIndex}. 885 * @return The codepoint as an int 886 * @deprecated ICU 3.2 887 * @obsolete ICU 3.2 888 */ 889 @Deprecated 890 public int setIndex(int index) { 891 setIndexOnly(index); 892 return current(); 893 } 894 895 /** 896 * Retrieve the index of the start of the input text. This is the begin 897 * index of the <tt>CharacterIterator</tt> or the start (i.e. 0) of the 898 * <tt>String</tt> over which this <tt>Normalizer</tt> is iterating 899 * @deprecated ICU 2.2. Use startIndex() instead. 900 * @return The codepoint as an int 901 * @see #startIndex 902 */ 903 @Deprecated 904 public int getBeginIndex() { 905 return 0; 906 } 907 908 /** 909 * Retrieve the index of the end of the input text. This is the end index 910 * of the <tt>CharacterIterator</tt> or the length of the <tt>String</tt> 911 * over which this <tt>Normalizer</tt> is iterating 912 * @deprecated ICU 2.2. Use endIndex() instead. 913 * @return The codepoint as an int 914 * @see #endIndex 915 */ 916 @Deprecated 917 public int getEndIndex() { 918 return endIndex(); 919 } 920 921 /** 922 * Retrieve the current iteration position in the input text that is 923 * being normalized. This method is useful in applications such as 924 * searching, where you need to be able to determine the position in 925 * the input text that corresponds to a given normalized output character. 926 * <p> 927 * <b>Note:</b> This method sets the position in the <em>input</em>, while 928 * {@link #next} and {@link #previous} iterate through characters in the 929 * <em>output</em>. This means that there is not necessarily a one-to-one 930 * correspondence between characters returned by <tt>next</tt> and 931 * <tt>previous</tt> and the indices passed to and returned from 932 * <tt>setIndex</tt> and {@link #getIndex}. 933 * @return The current iteration position 934 * @stable ICU 2.8 935 */ 936 public int getIndex() { 937 if(bufferPos<bufferLimit) { 938 return currentIndex; 939 } else { 940 return nextIndex; 941 } 942 } 943 944 /** 945 * Retrieve the index of the end of the input text-> This is the end index 946 * of the <tt>CharacterIterator</tt> or the length of the <tt>String</tt> 947 * over which this <tt>Normalizer</tt> is iterating 948 * @return The current iteration position 949 * @stable ICU 2.8 950 */ 951 public int endIndex() { 952 return text.getLength(); 953 } 954 955 //------------------------------------------------------------------------- 956 // Property access methods 957 //------------------------------------------------------------------------- 958 /** 959 * Set the normalization mode for this object. 960 * <p> 961 * <b>Note:</b>If the normalization mode is changed while iterating 962 * over a string, calls to {@link #next} and {@link #previous} may 963 * return previously buffers characters in the old normalization mode 964 * until the iteration is able to re-sync at the next base character. 965 * It is safest to call {@link #setText setText()}, {@link #first}, 966 * {@link #last}, etc. after calling <tt>setMode</tt>. 967 * <p> 968 * @param newMode the new mode for this <tt>Normalizer</tt>. 969 * The supported modes are: 970 * <ul> 971 * <li>{@link #COMPOSE} - Unicode canonical decompositiion 972 * followed by canonical composition. 973 * <li>{@link #COMPOSE_COMPAT} - Unicode compatibility decompositiion 974 * follwed by canonical composition. 975 * <li>{@link #DECOMP} - Unicode canonical decomposition 976 * <li>{@link #DECOMP_COMPAT} - Unicode compatibility decomposition. 977 * <li>{@link #NO_OP} - Do nothing but return characters 978 * from the underlying input text. 979 * </ul> 980 * 981 * @see #getMode 982 * @stable ICU 2.8 983 */ 984 public void setMode(Mode newMode) { 985 mode = newMode; 986 } 987 /** 988 * Return the basic operation performed by this <tt>Normalizer</tt> 989 * 990 * @see #setMode 991 * @stable ICU 2.8 992 */ 993 public Mode getMode() { 994 return mode; 995 } 996 997 /** 998 * Set the input text over which this <tt>Normalizer</tt> will iterate. 999 * The iteration position is set to the beginning of the input text-> 1000 * @param newText The new string to be normalized. 1001 * @stable ICU 2.8 1002 */ 1003 public void setText(String newText) { 1004 1005 UCharacterIterator newIter = UCharacterIterator.getInstance(newText); 1006 if (newIter == null) { 1007 throw new InternalError("Could not create a new UCharacterIterator"); 1008 } 1009 text = newIter; 1010 reset(); 1011 } 1012 1013 /** 1014 * Set the input text over which this <tt>Normalizer</tt> will iterate. 1015 * The iteration position is set to the beginning of the input text-> 1016 * @param newText The new string to be normalized. 1017 * @stable ICU 2.8 1018 */ 1019 public void setText(CharacterIterator newText) { 1020 1021 UCharacterIterator newIter = UCharacterIterator.getInstance(newText); 1022 if (newIter == null) { 1023 throw new InternalError("Could not create a new UCharacterIterator"); 1024 } 1025 text = newIter; 1026 currentIndex=nextIndex=0; 1027 clearBuffer(); 1028 } 1029 1030 //------------------------------------------------------------------------- 1031 // Private utility methods 1032 //------------------------------------------------------------------------- 1033 1034 1035 /* backward iteration --------------------------------------------------- */ 1036 1037 /* 1038 * read backwards and get norm32 1039 * return 0 if the character is <minC 1040 * if c2!=0 then (c2, c) is a surrogate pair (reversed - c2 is first 1041 * surrogate but read second!) 1042 */ 1043 1044 private static long getPrevNorm32(UCharacterIterator src, 1045 int/*unsigned*/ minC, 1046 int/*unsigned*/ mask, 1047 char[] chars) { 1048 long norm32; 1049 int ch=0; 1050 /* need src.hasPrevious() */ 1051 if((ch=src.previous()) == UCharacterIterator.DONE) { 1052 return 0; 1053 } 1054 chars[0]=(char)ch; 1055 chars[1]=0; 1056 1057 /* check for a surrogate before getting norm32 to see if we need to 1058 * predecrement further */ 1059 if(chars[0]<minC) { 1060 return 0; 1061 } else if(!UTF16.isSurrogate(chars[0])) { 1062 return NormalizerImpl.getNorm32(chars[0]); 1063 } else if(UTF16.isLeadSurrogate(chars[0]) || (src.getIndex()==0)) { 1064 /* unpaired surrogate */ 1065 chars[1]=(char)src.current(); 1066 return 0; 1067 } else if(UTF16.isLeadSurrogate(chars[1]=(char)src.previous())) { 1068 norm32=NormalizerImpl.getNorm32(chars[1]); 1069 if((norm32&mask)==0) { 1070 /* all surrogate pairs with this lead surrogate have irrelevant 1071 * data */ 1072 return 0; 1073 } else { 1074 /* norm32 must be a surrogate special */ 1075 return NormalizerImpl.getNorm32FromSurrogatePair(norm32,chars[0]); 1076 } 1077 } else { 1078 /* unpaired second surrogate, undo the c2=src.previous() movement */ 1079 src.moveIndex( 1); 1080 return 0; 1081 } 1082 } 1083 1084 private interface IsPrevBoundary{ 1085 public boolean isPrevBoundary(UCharacterIterator src, 1086 int/*unsigned*/ minC, 1087 int/*unsigned*/ mask, 1088 char[] chars); 1089 } 1090 private static final class IsPrevNFDSafe implements IsPrevBoundary{ 1091 /* 1092 * for NF*D: 1093 * read backwards and check if the lead combining class is 0 1094 * if c2!=0 then (c2, c) is a surrogate pair (reversed - c2 is first 1095 * surrogate but read second!) 1096 */ 1097 public boolean isPrevBoundary(UCharacterIterator src, 1098 int/*unsigned*/ minC, 1099 int/*unsigned*/ ccOrQCMask, 1100 char[] chars) { 1101 1102 return NormalizerImpl.isNFDSafe(getPrevNorm32(src, minC, 1103 ccOrQCMask, chars), 1104 ccOrQCMask, 1105 ccOrQCMask& NormalizerImpl.QC_MASK); 1106 } 1107 } 1108 1109 private static final class IsPrevTrueStarter implements IsPrevBoundary{ 1110 /* 1111 * read backwards and check if the character is (or its decomposition 1112 * begins with) a "true starter" (cc==0 and NF*C_YES) 1113 * if c2!=0 then (c2, c) is a surrogate pair (reversed - c2 is first 1114 * surrogate but read second!) 1115 */ 1116 public boolean isPrevBoundary(UCharacterIterator src, 1117 int/*unsigned*/ minC, 1118 int/*unsigned*/ ccOrQCMask, 1119 char[] chars) { 1120 long norm32; 1121 int/*unsigned*/ decompQCMask; 1122 1123 decompQCMask=(ccOrQCMask<<2)&0xf; /*decomposition quick check mask*/ 1124 norm32=getPrevNorm32(src, minC, ccOrQCMask|decompQCMask, chars); 1125 return NormalizerImpl.isTrueStarter(norm32,ccOrQCMask,decompQCMask); 1126 } 1127 } 1128 1129 private static int findPreviousIterationBoundary(UCharacterIterator src, 1130 IsPrevBoundary obj, 1131 int/*unsigned*/ minC, 1132 int/*mask*/ mask, 1133 char[] buffer, 1134 int[] startIndex) { 1135 char[] chars=new char[2]; 1136 boolean isBoundary; 1137 1138 /* fill the buffer from the end backwards */ 1139 startIndex[0] = buffer.length; 1140 chars[0]=0; 1141 while(src.getIndex()>0 && chars[0]!=UCharacterIterator.DONE) { 1142 isBoundary=obj.isPrevBoundary(src, minC, mask, chars); 1143 1144 /* always write this character to the front of the buffer */ 1145 /* make sure there is enough space in the buffer */ 1146 if(startIndex[0] < (chars[1]==0 ? 1 : 2)) { 1147 1148 // grow the buffer 1149 char[] newBuf = new char[buffer.length*2]; 1150 /* move the current buffer contents up */ 1151 System.arraycopy(buffer,startIndex[0],newBuf, 1152 newBuf.length-(buffer.length-startIndex[0]), 1153 buffer.length-startIndex[0]); 1154 //adjust the startIndex 1155 startIndex[0]+=newBuf.length-buffer.length; 1156 1157 buffer=newBuf; 1158 newBuf=null; 1159 1160 } 1161 1162 buffer[--startIndex[0]]=chars[0]; 1163 if(chars[1]!=0) { 1164 buffer[--startIndex[0]]=chars[1]; 1165 } 1166 1167 /* stop if this just-copied character is a boundary */ 1168 if(isBoundary) { 1169 break; 1170 } 1171 } 1172 1173 /* return the length of the buffer contents */ 1174 return buffer.length-startIndex[0]; 1175 } 1176 1177 private static int previous(UCharacterIterator src, 1178 char[] dest, int destStart, int destLimit, 1179 Mode mode, 1180 boolean doNormalize, 1181 boolean[] pNeededToNormalize, 1182 int options) { 1183 1184 IsPrevBoundary isPreviousBoundary; 1185 int destLength, bufferLength; 1186 int/*unsigned*/ mask; 1187 int c,c2; 1188 1189 char minC; 1190 int destCapacity = destLimit-destStart; 1191 destLength=0; 1192 1193 if(pNeededToNormalize!=null) { 1194 pNeededToNormalize[0]=false; 1195 } 1196 minC = (char)mode.getMinC(); 1197 mask = mode.getMask(); 1198 isPreviousBoundary = mode.getPrevBoundary(); 1199 1200 if(isPreviousBoundary==null) { 1201 destLength=0; 1202 if((c=src.previous())>=0) { 1203 destLength=1; 1204 if(UTF16.isTrailSurrogate((char)c)) { 1205 c2= src.previous(); 1206 if(c2!= UCharacterIterator.DONE) { 1207 if(UTF16.isLeadSurrogate((char)c2)) { 1208 if(destCapacity>=2) { 1209 dest[1]=(char)c; // trail surrogate 1210 destLength=2; 1211 } 1212 // lead surrogate to be written below 1213 c=c2; 1214 } else { 1215 src.moveIndex(1); 1216 } 1217 } 1218 } 1219 1220 if(destCapacity>0) { 1221 dest[0]=(char)c; 1222 } 1223 } 1224 return destLength; 1225 } 1226 1227 char[] buffer = new char[100]; 1228 int[] startIndex= new int[1]; 1229 bufferLength=findPreviousIterationBoundary(src, 1230 isPreviousBoundary, 1231 minC, mask,buffer, 1232 startIndex); 1233 if(bufferLength>0) { 1234 if(doNormalize) { 1235 destLength=NormalizerBase.normalize(buffer,startIndex[0], 1236 startIndex[0]+bufferLength, 1237 dest, destStart,destLimit, 1238 mode, options); 1239 1240 if(pNeededToNormalize!=null) { 1241 pNeededToNormalize[0]=destLength!=bufferLength || 1242 Utility.arrayRegionMatches( 1243 buffer,0,dest, 1244 destStart,destLimit 1245 ); 1246 } 1247 } else { 1248 /* just copy the source characters */ 1249 if(destCapacity>0) { 1250 System.arraycopy(buffer,startIndex[0],dest,0, 1251 (bufferLength<destCapacity) ? 1252 bufferLength : destCapacity 1253 ); 1254 } 1255 } 1256 } 1257 1258 1259 return destLength; 1260 } 1261 1262 1263 1264 /* forward iteration ---------------------------------------------------- */ 1265 /* 1266 * read forward and check if the character is a next-iteration boundary 1267 * if c2!=0 then (c, c2) is a surrogate pair 1268 */ 1269 private interface IsNextBoundary{ 1270 boolean isNextBoundary(UCharacterIterator src, 1271 int/*unsigned*/ minC, 1272 int/*unsigned*/ mask, 1273 int[] chars); 1274 } 1275 /* 1276 * read forward and get norm32 1277 * return 0 if the character is <minC 1278 * if c2!=0 then (c2, c) is a surrogate pair 1279 * always reads complete characters 1280 */ 1281 private static long /*unsigned*/ getNextNorm32(UCharacterIterator src, 1282 int/*unsigned*/ minC, 1283 int/*unsigned*/ mask, 1284 int[] chars) { 1285 long norm32; 1286 1287 /* need src.hasNext() to be true */ 1288 chars[0]=src.next(); 1289 chars[1]=0; 1290 1291 if(chars[0]<minC) { 1292 return 0; 1293 } 1294 1295 norm32=NormalizerImpl.getNorm32((char)chars[0]); 1296 if(UTF16.isLeadSurrogate((char)chars[0])) { 1297 if(src.current()!=UCharacterIterator.DONE && 1298 UTF16.isTrailSurrogate((char)(chars[1]=src.current()))) { 1299 src.moveIndex(1); /* skip the c2 surrogate */ 1300 if((norm32&mask)==0) { 1301 /* irrelevant data */ 1302 return 0; 1303 } else { 1304 /* norm32 must be a surrogate special */ 1305 return NormalizerImpl.getNorm32FromSurrogatePair(norm32,(char)chars[1]); 1306 } 1307 } else { 1308 /* unmatched surrogate */ 1309 return 0; 1310 } 1311 } 1312 return norm32; 1313 } 1314 1315 1316 /* 1317 * for NF*D: 1318 * read forward and check if the lead combining class is 0 1319 * if c2!=0 then (c, c2) is a surrogate pair 1320 */ 1321 private static final class IsNextNFDSafe implements IsNextBoundary{ 1322 public boolean isNextBoundary(UCharacterIterator src, 1323 int/*unsigned*/ minC, 1324 int/*unsigned*/ ccOrQCMask, 1325 int[] chars) { 1326 return NormalizerImpl.isNFDSafe(getNextNorm32(src,minC,ccOrQCMask,chars), 1327 ccOrQCMask, ccOrQCMask&NormalizerImpl.QC_MASK); 1328 } 1329 } 1330 1331 /* 1332 * for NF*C: 1333 * read forward and check if the character is (or its decomposition begins 1334 * with) a "true starter" (cc==0 and NF*C_YES) 1335 * if c2!=0 then (c, c2) is a surrogate pair 1336 */ 1337 private static final class IsNextTrueStarter implements IsNextBoundary{ 1338 public boolean isNextBoundary(UCharacterIterator src, 1339 int/*unsigned*/ minC, 1340 int/*unsigned*/ ccOrQCMask, 1341 int[] chars) { 1342 long norm32; 1343 int/*unsigned*/ decompQCMask; 1344 1345 decompQCMask=(ccOrQCMask<<2)&0xf; /*decomposition quick check mask*/ 1346 norm32=getNextNorm32(src, minC, ccOrQCMask|decompQCMask, chars); 1347 return NormalizerImpl.isTrueStarter(norm32, ccOrQCMask, decompQCMask); 1348 } 1349 } 1350 1351 private static int findNextIterationBoundary(UCharacterIterator src, 1352 IsNextBoundary obj, 1353 int/*unsigned*/ minC, 1354 int/*unsigned*/ mask, 1355 char[] buffer) { 1356 if(src.current()==UCharacterIterator.DONE) { 1357 return 0; 1358 } 1359 1360 /* get one character and ignore its properties */ 1361 int[] chars = new int[2]; 1362 chars[0]=src.next(); 1363 buffer[0]=(char)chars[0]; 1364 int bufferIndex = 1; 1365 1366 if(UTF16.isLeadSurrogate((char)chars[0])&& 1367 src.current()!=UCharacterIterator.DONE) { 1368 if(UTF16.isTrailSurrogate((char)(chars[1]=src.next()))) { 1369 buffer[bufferIndex++]=(char)chars[1]; 1370 } else { 1371 src.moveIndex(-1); /* back out the non-trail-surrogate */ 1372 } 1373 } 1374 1375 /* get all following characters until we see a boundary */ 1376 /* checking hasNext() instead of c!=DONE on the off-chance that U+ffff 1377 * is part of the string */ 1378 while( src.current()!=UCharacterIterator.DONE) { 1379 if(obj.isNextBoundary(src, minC, mask, chars)) { 1380 /* back out the latest movement to stop at the boundary */ 1381 src.moveIndex(chars[1]==0 ? -1 : -2); 1382 break; 1383 } else { 1384 if(bufferIndex+(chars[1]==0 ? 1 : 2)<=buffer.length) { 1385 buffer[bufferIndex++]=(char)chars[0]; 1386 if(chars[1]!=0) { 1387 buffer[bufferIndex++]=(char)chars[1]; 1388 } 1389 } else { 1390 char[] newBuf = new char[buffer.length*2]; 1391 System.arraycopy(buffer,0,newBuf,0,bufferIndex); 1392 buffer = newBuf; 1393 buffer[bufferIndex++]=(char)chars[0]; 1394 if(chars[1]!=0) { 1395 buffer[bufferIndex++]=(char)chars[1]; 1396 } 1397 } 1398 } 1399 } 1400 1401 /* return the length of the buffer contents */ 1402 return bufferIndex; 1403 } 1404 1405 private static int next(UCharacterIterator src, 1406 char[] dest, int destStart, int destLimit, 1407 NormalizerBase.Mode mode, 1408 boolean doNormalize, 1409 boolean[] pNeededToNormalize, 1410 int options) { 1411 1412 IsNextBoundary isNextBoundary; 1413 int /*unsigned*/ mask; 1414 int /*unsigned*/ bufferLength; 1415 int c,c2; 1416 char minC; 1417 int destCapacity = destLimit - destStart; 1418 int destLength = 0; 1419 if(pNeededToNormalize!=null) { 1420 pNeededToNormalize[0]=false; 1421 } 1422 1423 minC = (char)mode.getMinC(); 1424 mask = mode.getMask(); 1425 isNextBoundary = mode.getNextBoundary(); 1426 1427 if(isNextBoundary==null) { 1428 destLength=0; 1429 c=src.next(); 1430 if(c!=UCharacterIterator.DONE) { 1431 destLength=1; 1432 if(UTF16.isLeadSurrogate((char)c)) { 1433 c2= src.next(); 1434 if(c2!= UCharacterIterator.DONE) { 1435 if(UTF16.isTrailSurrogate((char)c2)) { 1436 if(destCapacity>=2) { 1437 dest[1]=(char)c2; // trail surrogate 1438 destLength=2; 1439 } 1440 // lead surrogate to be written below 1441 } else { 1442 src.moveIndex(-1); 1443 } 1444 } 1445 } 1446 1447 if(destCapacity>0) { 1448 dest[0]=(char)c; 1449 } 1450 } 1451 return destLength; 1452 } 1453 1454 char[] buffer=new char[100]; 1455 int[] startIndex = new int[1]; 1456 bufferLength=findNextIterationBoundary(src,isNextBoundary, minC, mask, 1457 buffer); 1458 if(bufferLength>0) { 1459 if(doNormalize) { 1460 destLength=mode.normalize(buffer,startIndex[0],bufferLength, 1461 dest,destStart,destLimit, options); 1462 1463 if(pNeededToNormalize!=null) { 1464 pNeededToNormalize[0]=destLength!=bufferLength || 1465 Utility.arrayRegionMatches(buffer,startIndex[0], 1466 dest,destStart, 1467 destLength); 1468 } 1469 } else { 1470 /* just copy the source characters */ 1471 if(destCapacity>0) { 1472 System.arraycopy(buffer,0,dest,destStart, 1473 Math.min(bufferLength,destCapacity) 1474 ); 1475 } 1476 1477 1478 } 1479 } 1480 return destLength; 1481 } 1482 1483 private void clearBuffer() { 1484 bufferLimit=bufferStart=bufferPos=0; 1485 } 1486 1487 private boolean nextNormalize() { 1488 1489 clearBuffer(); 1490 currentIndex=nextIndex; 1491 text.setIndex(nextIndex); 1492 1493 bufferLimit=next(text,buffer,bufferStart,buffer.length,mode,true,null,options); 1494 1495 nextIndex=text.getIndex(); 1496 return (bufferLimit>0); 1497 } 1498 1499 private boolean previousNormalize() { 1500 1501 clearBuffer(); 1502 nextIndex=currentIndex; 1503 text.setIndex(currentIndex); 1504 bufferLimit=previous(text,buffer,bufferStart,buffer.length,mode,true,null,options); 1505 1506 currentIndex=text.getIndex(); 1507 bufferPos = bufferLimit; 1508 return bufferLimit>0; 1509 } 1510 1511 private int getCodePointAt(int index) { 1512 if( UTF16.isSurrogate(buffer[index])) { 1513 if(UTF16.isLeadSurrogate(buffer[index])) { 1514 if((index+1)<bufferLimit && 1515 UTF16.isTrailSurrogate(buffer[index+1])) { 1516 return UCharacterProperty.getRawSupplementary( 1517 buffer[index], 1518 buffer[index+1] 1519 ); 1520 } 1521 }else if(UTF16.isTrailSurrogate(buffer[index])) { 1522 if(index>0 && UTF16.isLeadSurrogate(buffer[index-1])) { 1523 return UCharacterProperty.getRawSupplementary( 1524 buffer[index-1], 1525 buffer[index] 1526 ); 1527 } 1528 } 1529 } 1530 return buffer[index]; 1531 1532 } 1533 1534 /** 1535 * Internal API 1536 * @internal 1537 */ 1538 public static boolean isNFSkippable(int c, Mode mode) { 1539 return mode.isNFSkippable(c); 1540 } 1541 1542 // 1543 // Options 1544 // 1545 1546 /* 1547 * Default option for Unicode 3.2.0 normalization. 1548 * Corrigendum 4 was fixed in Unicode 3.2.0 but isn't supported in 1549 * IDNA/StringPrep. 1550 * The public review issue #29 was fixed in Unicode 4.1.0. Corrigendum 5 1551 * allowed Unicode 3.2 to 4.0.1 to apply the fix for PRI #29, but it isn't 1552 * supported by IDNA/StringPrep as well as Corrigendum 4. 1553 */ 1554 public static final int UNICODE_3_2_0_ORIGINAL = 1555 UNICODE_3_2 | 1556 NormalizerImpl.WITHOUT_CORRIGENDUM4_CORRECTIONS | 1557 NormalizerImpl.BEFORE_PRI_29; 1558 1559 /* 1560 * Default option for the latest Unicode normalization. This option is 1561 * provided mainly for testing. 1562 * The value zero means that normalization is done with the fixes for 1563 * - Corrigendum 4 (Five CJK Canonical Mapping Errors) 1564 * - Corrigendum 5 (Normalization Idempotency) 1565 */ 1566 public static final int UNICODE_LATEST = 0x00; 1567 1568 // 1569 // public constructor and methods for java.text.Normalizer and 1570 // sun.text.Normalizer 1571 // 1572 1573 /** 1574 * Creates a new <tt>Normalizer</tt> object for iterating over the 1575 * normalized form of a given string. 1576 * 1577 * @param str The string to be normalized. The normalization 1578 * will start at the beginning of the string. 1579 * 1580 * @param mode The normalization mode. 1581 */ 1582 public NormalizerBase(String str, Mode mode) { 1583 this(str, mode, UNICODE_LATEST); 1584 } 1585 1586 /** 1587 * Normalizes a <code>String</code> using the given normalization form. 1588 * 1589 * @param str the input string to be normalized. 1590 * @param form the normalization form 1591 */ 1592 public static String normalize(String str, Normalizer.Form form) { 1593 return normalize(str, form, UNICODE_LATEST); 1594 } 1595 1596 /** 1597 * Normalizes a <code>String</code> using the given normalization form. 1598 * 1599 * @param str the input string to be normalized. 1600 * @param form the normalization form 1601 * @param options the optional features to be enabled. 1602 */ 1603 public static String normalize(String str, Normalizer.Form form, int options) { 1604 int len = str.length(); 1605 boolean asciiOnly = true; 1606 if (len < 80) { 1607 for (int i = 0; i < len; i++) { 1608 if (str.charAt(i) > 127) { 1609 asciiOnly = false; 1610 break; 1611 } 1612 } 1613 } else { 1614 char[] a = str.toCharArray(); 1615 for (int i = 0; i < len; i++) { 1616 if (a[i] > 127) { 1617 asciiOnly = false; 1618 break; 1619 } 1620 } 1621 } 1622 1623 switch (form) { 1624 case NFC : 1625 return asciiOnly ? str : NFC.normalize(str, options); 1626 case NFD : 1627 return asciiOnly ? str : NFD.normalize(str, options); 1628 case NFKC : 1629 return asciiOnly ? str : NFKC.normalize(str, options); 1630 case NFKD : 1631 return asciiOnly ? str : NFKD.normalize(str, options); 1632 } 1633 1634 throw new IllegalArgumentException("Unexpected normalization form: " + 1635 form); 1636 } 1637 1638 /** 1639 * Test if a string is in a given normalization form. 1640 * This is semantically equivalent to source.equals(normalize(source, mode)). 1641 * 1642 * Unlike quickCheck(), this function returns a definitive result, 1643 * never a "maybe". 1644 * For NFD, NFKD, and FCD, both functions work exactly the same. 1645 * For NFC and NFKC where quickCheck may return "maybe", this function will 1646 * perform further tests to arrive at a true/false result. 1647 * @param str the input string to be checked to see if it is normalized 1648 * @param form the normalization form 1649 * @param options the optional features to be enabled. 1650 */ 1651 public static boolean isNormalized(String str, Normalizer.Form form) { 1652 return isNormalized(str, form, UNICODE_LATEST); 1653 } 1654 1655 /** 1656 * Test if a string is in a given normalization form. 1657 * This is semantically equivalent to source.equals(normalize(source, mode)). 1658 * 1659 * Unlike quickCheck(), this function returns a definitive result, 1660 * never a "maybe". 1661 * For NFD, NFKD, and FCD, both functions work exactly the same. 1662 * For NFC and NFKC where quickCheck may return "maybe", this function will 1663 * perform further tests to arrive at a true/false result. 1664 * @param str the input string to be checked to see if it is normalized 1665 * @param form the normalization form 1666 * @param options the optional features to be enabled. 1667 */ 1668 public static boolean isNormalized(String str, Normalizer.Form form, int options) { 1669 switch (form) { 1670 case NFC: 1671 return (NFC.quickCheck(str.toCharArray(),0,str.length(),false,NormalizerImpl.getNX(options))==YES); 1672 case NFD: 1673 return (NFD.quickCheck(str.toCharArray(),0,str.length(),false,NormalizerImpl.getNX(options))==YES); 1674 case NFKC: 1675 return (NFKC.quickCheck(str.toCharArray(),0,str.length(),false,NormalizerImpl.getNX(options))==YES); 1676 case NFKD: 1677 return (NFKD.quickCheck(str.toCharArray(),0,str.length(),false,NormalizerImpl.getNX(options))==YES); 1678 } 1679 1680 throw new IllegalArgumentException("Unexpected normalization form: " + 1681 form); 1682 } 1683 }