icuSources/common/unicode/normalizer2.h

   1 // © 2016 and later: Unicode, Inc. and others.
   2 // License & terms of use: http://www.unicode.org/copyright.html
   3 /*
   4 *******************************************************************************
   5 *
   6 *   Copyright (C) 2009-2013, International Business Machines
   7 *   Corporation and others.  All Rights Reserved.
   8 *
   9 *******************************************************************************
  10 *   file name:  normalizer2.h
  11 *   encoding:   UTF-8
  12 *   tab size:   8 (not used)
  13 *   indentation:4
  14 *
  15 *   created on: 2009nov22
  16 *   created by: Markus W. Scherer
  17 */
  18
  19 #ifndef __NORMALIZER2_H__
  20 #define __NORMALIZER2_H__
  21
  22 /**
  23  * \file
  24  * \brief C++ API: New API for Unicode Normalization.
  25  */
  26
  27 #include "unicode/utypes.h"
  28
  29 #if !UCONFIG_NO_NORMALIZATION
  30
  31 #include "unicode/stringpiece.h"
  32 #include "unicode/uniset.h"
  33 #include "unicode/unistr.h"
  34 #include "unicode/unorm2.h"
  35
  36 #if U_SHOW_CPLUSPLUS_API
  37 U_NAMESPACE_BEGIN
  38
  39 class ByteSink;
  40
  41 /**
  42  * Unicode normalization functionality for standard Unicode normalization or
  43  * for using custom mapping tables.
  44  * All instances of this class are unmodifiable/immutable.
  45  * Instances returned by getInstance() are singletons that must not be deleted by the caller.
  46  * The Normalizer2 class is not intended for public subclassing.
  47  *
  48  * The primary functions are to produce a normalized string and to detect whether
  49  * a string is already normalized.
  50  * The most commonly used normalization forms are those defined in
  51  * http://www.unicode.org/unicode/reports/tr15/
  52  * However, this API supports additional normalization forms for specialized purposes.
  53  * For example, NFKC_Casefold is provided via getInstance("nfkc_cf", COMPOSE)
  54  * and can be used in implementations of UTS #46.
  55  *
  56  * Not only are the standard compose and decompose modes supplied,
  57  * but additional modes are provided as documented in the Mode enum.
  58  *
  59  * Some of the functions in this class identify normalization boundaries.
  60  * At a normalization boundary, the portions of the string
  61  * before it and starting from it do not interact and can be handled independently.
  62  *
  63  * The spanQuickCheckYes() stops at a normalization boundary.
  64  * When the goal is a normalized string, then the text before the boundary
  65  * can be copied, and the remainder can be processed with normalizeSecondAndAppend().
  66  *
  67  * The hasBoundaryBefore(), hasBoundaryAfter() and isInert() functions test whether
  68  * a character is guaranteed to be at a normalization boundary,
  69  * regardless of context.
  70  * This is used for moving from one normalization boundary to the next
  71  * or preceding boundary, and for performing iterative normalization.
  72  *
  73  * Iterative normalization is useful when only a small portion of a
  74  * longer string needs to be processed.
  75  * For example, in ICU, iterative normalization is used by the NormalizationTransliterator
  76  * (to avoid replacing already-normalized text) and ucol_nextSortKeyPart()
  77  * (to process only the substring for which sort key bytes are computed).
  78  *
  79  * The set of normalization boundaries returned by these functions may not be
  80  * complete: There may be more boundaries that could be returned.
  81  * Different functions may return different boundaries.
  82  * @stable ICU 4.4
  83  */
  84 class U_COMMON_API Normalizer2 : public UObject {
  85 public:
  86     /**
  87      * Destructor.
  88      * @stable ICU 4.4
  89      */
  90     ~Normalizer2();
  91
  92     /**
  93      * Returns a Normalizer2 instance for Unicode NFC normalization.
  94      * Same as getInstance(NULL, "nfc", UNORM2_COMPOSE, errorCode).
  95      * Returns an unmodifiable singleton instance. Do not delete it.
  96      * @param errorCode Standard ICU error code. Its input value must
  97      *                  pass the U_SUCCESS() test, or else the function returns
  98      *                  immediately. Check for U_FAILURE() on output or use with
  99      *                  function chaining. (See User Guide for details.)
 100      * @return the requested Normalizer2, if successful
 101      * @stable ICU 49
 102      */
 103     static const Normalizer2 *
 104     getNFCInstance(UErrorCode &errorCode);
 105
 106     /**
 107      * Returns a Normalizer2 instance for Unicode NFD normalization.
 108      * Same as getInstance(NULL, "nfc", UNORM2_DECOMPOSE, errorCode).
 109      * Returns an unmodifiable singleton instance. Do not delete it.
 110      * @param errorCode Standard ICU error code. Its input value must
 111      *                  pass the U_SUCCESS() test, or else the function returns
 112      *                  immediately. Check for U_FAILURE() on output or use with
 113      *                  function chaining. (See User Guide for details.)
 114      * @return the requested Normalizer2, if successful
 115      * @stable ICU 49
 116      */
 117     static const Normalizer2 *
 118     getNFDInstance(UErrorCode &errorCode);
 119
 120     /**
 121      * Returns a Normalizer2 instance for Unicode NFKC normalization.
 122      * Same as getInstance(NULL, "nfkc", UNORM2_COMPOSE, errorCode).
 123      * Returns an unmodifiable singleton instance. Do not delete it.
 124      * @param errorCode Standard ICU error code. Its input value must
 125      *                  pass the U_SUCCESS() test, or else the function returns
 126      *                  immediately. Check for U_FAILURE() on output or use with
 127      *                  function chaining. (See User Guide for details.)
 128      * @return the requested Normalizer2, if successful
 129      * @stable ICU 49
 130      */
 131     static const Normalizer2 *
 132     getNFKCInstance(UErrorCode &errorCode);
 133
 134     /**
 135      * Returns a Normalizer2 instance for Unicode NFKD normalization.
 136      * Same as getInstance(NULL, "nfkc", UNORM2_DECOMPOSE, errorCode).
 137      * Returns an unmodifiable singleton instance. Do not delete it.
 138      * @param errorCode Standard ICU error code. Its input value must
 139      *                  pass the U_SUCCESS() test, or else the function returns
 140      *                  immediately. Check for U_FAILURE() on output or use with
 141      *                  function chaining. (See User Guide for details.)
 142      * @return the requested Normalizer2, if successful
 143      * @stable ICU 49
 144      */
 145     static const Normalizer2 *
 146     getNFKDInstance(UErrorCode &errorCode);
 147
 148     /**
 149      * Returns a Normalizer2 instance for Unicode NFKC_Casefold normalization.
 150      * Same as getInstance(NULL, "nfkc_cf", UNORM2_COMPOSE, errorCode).
 151      * Returns an unmodifiable singleton instance. Do not delete it.
 152      * @param errorCode Standard ICU error code. Its input value must
 153      *                  pass the U_SUCCESS() test, or else the function returns
 154      *                  immediately. Check for U_FAILURE() on output or use with
 155      *                  function chaining. (See User Guide for details.)
 156      * @return the requested Normalizer2, if successful
 157      * @stable ICU 49
 158      */
 159     static const Normalizer2 *
 160     getNFKCCasefoldInstance(UErrorCode &errorCode);
 161
 162     /**
 163      * Returns a Normalizer2 instance which uses the specified data file
 164      * (packageName/name similar to ucnv_openPackage() and ures_open()/ResourceBundle)
 165      * and which composes or decomposes text according to the specified mode.
 166      * Returns an unmodifiable singleton instance. Do not delete it.
 167      *
 168      * Use packageName=NULL for data files that are part of ICU's own data.
 169      * Use name="nfc" and UNORM2_COMPOSE/UNORM2_DECOMPOSE for Unicode standard NFC/NFD.
 170      * Use name="nfkc" and UNORM2_COMPOSE/UNORM2_DECOMPOSE for Unicode standard NFKC/NFKD.
 171      * Use name="nfkc_cf" and UNORM2_COMPOSE for Unicode standard NFKC_CF=NFKC_Casefold.
 172      *
 173      * @param packageName NULL for ICU built-in data, otherwise application data package name
 174      * @param name "nfc" or "nfkc" or "nfkc_cf" or name of custom data file
 175      * @param mode normalization mode (compose or decompose etc.)
 176      * @param errorCode Standard ICU error code. Its input value must
 177      *                  pass the U_SUCCESS() test, or else the function returns
 178      *                  immediately. Check for U_FAILURE() on output or use with
 179      *                  function chaining. (See User Guide for details.)
 180      * @return the requested Normalizer2, if successful
 181      * @stable ICU 4.4
 182      */
 183     static const Normalizer2 *
 184     getInstance(const char *packageName,
 185                 const char *name,
 186                 UNormalization2Mode mode,
 187                 UErrorCode &errorCode);
 188
 189     /**
 190      * Returns the normalized form of the source string.
 191      * @param src source string
 192      * @param errorCode Standard ICU error code. Its input value must
 193      *                  pass the U_SUCCESS() test, or else the function returns
 194      *                  immediately. Check for U_FAILURE() on output or use with
 195      *                  function chaining. (See User Guide for details.)
 196      * @return normalized src
 197      * @stable ICU 4.4
 198      */
 199     UnicodeString
 200     normalize(const UnicodeString &src, UErrorCode &errorCode) const {
 201         UnicodeString result;
 202         normalize(src, result, errorCode);
 203         return result;
 204     }
 205     /**
 206      * Writes the normalized form of the source string to the destination string
 207      * (replacing its contents) and returns the destination string.
 208      * The source and destination strings must be different objects.
 209      * @param src source string
 210      * @param dest destination string; its contents is replaced with normalized src
 211      * @param errorCode Standard ICU error code. Its input value must
 212      *                  pass the U_SUCCESS() test, or else the function returns
 213      *                  immediately. Check for U_FAILURE() on output or use with
 214      *                  function chaining. (See User Guide for details.)
 215      * @return dest
 216      * @stable ICU 4.4
 217      */
 218     virtual UnicodeString &
 219     normalize(const UnicodeString &src,
 220               UnicodeString &dest,
 221               UErrorCode &errorCode) const = 0;
 222
 223     /**
 224      * Normalizes a UTF-8 string and optionally records how source substrings
 225      * relate to changed and unchanged result substrings.
 226      *
 227      * Currently implemented completely only for "compose" modes,
 228      * such as for NFC, NFKC, and NFKC_Casefold
 229      * (UNORM2_COMPOSE and UNORM2_COMPOSE_CONTIGUOUS).
 230      * Otherwise currently converts to & from UTF-16 and does not support edits.
 231      *
 232      * @param options   Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT and U_EDITS_NO_RESET.
 233      * @param src       Source UTF-8 string.
 234      * @param sink      A ByteSink to which the normalized UTF-8 result string is written.
 235      *                  sink.Flush() is called at the end.
 236      * @param edits     Records edits for index mapping, working with styled text,
 237      *                  and getting only changes (if any).
 238      *                  The Edits contents is undefined if any error occurs.
 239      *                  This function calls edits->reset() first unless
 240      *                  options includes U_EDITS_NO_RESET. edits can be nullptr.
 241      * @param errorCode Standard ICU error code. Its input value must
 242      *                  pass the U_SUCCESS() test, or else the function returns
 243      *                  immediately. Check for U_FAILURE() on output or use with
 244      *                  function chaining. (See User Guide for details.)
 245      * @draft ICU 60
 246      */
 247     virtual void
 248     normalizeUTF8(uint32_t options, StringPiece src, ByteSink &sink,
 249                   Edits *edits, UErrorCode &errorCode) const;
 250
 251     /**
 252      * Appends the normalized form of the second string to the first string
 253      * (merging them at the boundary) and returns the first string.
 254      * The result is normalized if the first string was normalized.
 255      * The first and second strings must be different objects.
 256      * @param first string, should be normalized
 257      * @param second string, will be normalized
 258      * @param errorCode Standard ICU error code. Its input value must
 259      *                  pass the U_SUCCESS() test, or else the function returns
 260      *                  immediately. Check for U_FAILURE() on output or use with
 261      *                  function chaining. (See User Guide for details.)
 262      * @return first
 263      * @stable ICU 4.4
 264      */
 265     virtual UnicodeString &
 266     normalizeSecondAndAppend(UnicodeString &first,
 267                              const UnicodeString &second,
 268                              UErrorCode &errorCode) const = 0;
 269     /**
 270      * Appends the second string to the first string
 271      * (merging them at the boundary) and returns the first string.
 272      * The result is normalized if both the strings were normalized.
 273      * The first and second strings must be different objects.
 274      * @param first string, should be normalized
 275      * @param second string, should be normalized
 276      * @param errorCode Standard ICU error code. Its input value must
 277      *                  pass the U_SUCCESS() test, or else the function returns
 278      *                  immediately. Check for U_FAILURE() on output or use with
 279      *                  function chaining. (See User Guide for details.)
 280      * @return first
 281      * @stable ICU 4.4
 282      */
 283     virtual UnicodeString &
 284     append(UnicodeString &first,
 285            const UnicodeString &second,
 286            UErrorCode &errorCode) const = 0;
 287
 288     /**
 289      * Gets the decomposition mapping of c.
 290      * Roughly equivalent to normalizing the String form of c
 291      * on a UNORM2_DECOMPOSE Normalizer2 instance, but much faster, and except that this function
 292      * returns FALSE and does not write a string
 293      * if c does not have a decomposition mapping in this instance's data.
 294      * This function is independent of the mode of the Normalizer2.
 295      * @param c code point
 296      * @param decomposition String object which will be set to c's
 297      *                      decomposition mapping, if there is one.
 298      * @return TRUE if c has a decomposition, otherwise FALSE
 299      * @stable ICU 4.6
 300      */
 301     virtual UBool
 302     getDecomposition(UChar32 c, UnicodeString &decomposition) const = 0;
 303
 304     /**
 305      * Gets the raw decomposition mapping of c.
 306      *
 307      * This is similar to the getDecomposition() method but returns the
 308      * raw decomposition mapping as specified in UnicodeData.txt or
 309      * (for custom data) in the mapping files processed by the gennorm2 tool.
 310      * By contrast, getDecomposition() returns the processed,
 311      * recursively-decomposed version of this mapping.
 312      *
 313      * When used on a standard NFKC Normalizer2 instance,
 314      * getRawDecomposition() returns the Unicode Decomposition_Mapping (dm) property.
 315      *
 316      * When used on a standard NFC Normalizer2 instance,
 317      * it returns the Decomposition_Mapping only if the Decomposition_Type (dt) is Canonical (Can);
 318      * in this case, the result contains either one or two code points (=1..4 char16_ts).
 319      *
 320      * This function is independent of the mode of the Normalizer2.
 321      * The default implementation returns FALSE.
 322      * @param c code point
 323      * @param decomposition String object which will be set to c's
 324      *                      raw decomposition mapping, if there is one.
 325      * @return TRUE if c has a decomposition, otherwise FALSE
 326      * @stable ICU 49
 327      */
 328     virtual UBool
 329     getRawDecomposition(UChar32 c, UnicodeString &decomposition) const;
 330
 331     /**
 332      * Performs pairwise composition of a & b and returns the composite if there is one.
 333      *
 334      * Returns a composite code point c only if c has a two-way mapping to a+b.
 335      * In standard Unicode normalization, this means that
 336      * c has a canonical decomposition to a+b
 337      * and c does not have the Full_Composition_Exclusion property.
 338      *
 339      * This function is independent of the mode of the Normalizer2.
 340      * The default implementation returns a negative value.
 341      * @param a A (normalization starter) code point.
 342      * @param b Another code point.
 343      * @return The non-negative composite code point if there is one; otherwise a negative value.
 344      * @stable ICU 49
 345      */
 346     virtual UChar32
 347     composePair(UChar32 a, UChar32 b) const;
 348
 349     /**
 350      * Gets the combining class of c.
 351      * The default implementation returns 0
 352      * but all standard implementations return the Unicode Canonical_Combining_Class value.
 353      * @param c code point
 354      * @return c's combining class
 355      * @stable ICU 49
 356      */
 357     virtual uint8_t
 358     getCombiningClass(UChar32 c) const;
 359
 360     /**
 361      * Tests if the string is normalized.
 362      * Internally, in cases where the quickCheck() method would return "maybe"
 363      * (which is only possible for the two COMPOSE modes) this method
 364      * resolves to "yes" or "no" to provide a definitive result,
 365      * at the cost of doing more work in those cases.
 366      * @param s input string
 367      * @param errorCode Standard ICU error code. Its input value must
 368      *                  pass the U_SUCCESS() test, or else the function returns
 369      *                  immediately. Check for U_FAILURE() on output or use with
 370      *                  function chaining. (See User Guide for details.)
 371      * @return TRUE if s is normalized
 372      * @stable ICU 4.4
 373      */
 374     virtual UBool
 375     isNormalized(const UnicodeString &s, UErrorCode &errorCode) const = 0;
 376     /**
 377      * Tests if the UTF-8 string is normalized.
 378      * Internally, in cases where the quickCheck() method would return "maybe"
 379      * (which is only possible for the two COMPOSE modes) this method
 380      * resolves to "yes" or "no" to provide a definitive result,
 381      * at the cost of doing more work in those cases.
 382      *
 383      * This works for all normalization modes,
 384      * but it is currently optimized for UTF-8 only for "compose" modes,
 385      * such as for NFC, NFKC, and NFKC_Casefold
 386      * (UNORM2_COMPOSE and UNORM2_COMPOSE_CONTIGUOUS).
 387      * For other modes it currently converts to UTF-16 and calls isNormalized().
 388      *
 389      * @param s UTF-8 input string
 390      * @param errorCode Standard ICU error code. Its input value must
 391      *                  pass the U_SUCCESS() test, or else the function returns
 392      *                  immediately. Check for U_FAILURE() on output or use with
 393      *                  function chaining. (See User Guide for details.)
 394      * @return TRUE if s is normalized
 395      * @draft ICU 60
 396      */
 397     virtual UBool
 398     isNormalizedUTF8(StringPiece s, UErrorCode &errorCode) const;
 399
 400
 401     /**
 402      * Tests if the string is normalized.
 403      * For the two COMPOSE modes, the result could be "maybe" in cases that
 404      * would take a little more work to resolve definitively.
 405      * Use spanQuickCheckYes() and normalizeSecondAndAppend() for a faster
 406      * combination of quick check + normalization, to avoid
 407      * re-checking the "yes" prefix.
 408      * @param s input string
 409      * @param errorCode Standard ICU error code. Its input value must
 410      *                  pass the U_SUCCESS() test, or else the function returns
 411      *                  immediately. Check for U_FAILURE() on output or use with
 412      *                  function chaining. (See User Guide for details.)
 413      * @return UNormalizationCheckResult
 414      * @stable ICU 4.4
 415      */
 416     virtual UNormalizationCheckResult
 417     quickCheck(const UnicodeString &s, UErrorCode &errorCode) const = 0;
 418
 419     /**
 420      * Returns the end of the normalized substring of the input string.
 421      * In other words, with <code>end=spanQuickCheckYes(s, ec);</code>
 422      * the substring <code>UnicodeString(s, 0, end)</code>
 423      * will pass the quick check with a "yes" result.
 424      *
 425      * The returned end index is usually one or more characters before the
 426      * "no" or "maybe" character: The end index is at a normalization boundary.
 427      * (See the class documentation for more about normalization boundaries.)
 428      *
 429      * When the goal is a normalized string and most input strings are expected
 430      * to be normalized already, then call this method,
 431      * and if it returns a prefix shorter than the input string,
 432      * copy that prefix and use normalizeSecondAndAppend() for the remainder.
 433      * @param s input string
 434      * @param errorCode Standard ICU error code. Its input value must
 435      *                  pass the U_SUCCESS() test, or else the function returns
 436      *                  immediately. Check for U_FAILURE() on output or use with
 437      *                  function chaining. (See User Guide for details.)
 438      * @return "yes" span end index
 439      * @stable ICU 4.4
 440      */
 441     virtual int32_t
 442     spanQuickCheckYes(const UnicodeString &s, UErrorCode &errorCode) const = 0;
 443
 444     /**
 445      * Tests if the character always has a normalization boundary before it,
 446      * regardless of context.
 447      * If true, then the character does not normalization-interact with
 448      * preceding characters.
 449      * In other words, a string containing this character can be normalized
 450      * by processing portions before this character and starting from this
 451      * character independently.
 452      * This is used for iterative normalization. See the class documentation for details.
 453      * @param c character to test
 454      * @return TRUE if c has a normalization boundary before it
 455      * @stable ICU 4.4
 456      */
 457     virtual UBool hasBoundaryBefore(UChar32 c) const = 0;
 458
 459     /**
 460      * Tests if the character always has a normalization boundary after it,
 461      * regardless of context.
 462      * If true, then the character does not normalization-interact with
 463      * following characters.
 464      * In other words, a string containing this character can be normalized
 465      * by processing portions up to this character and after this
 466      * character independently.
 467      * This is used for iterative normalization. See the class documentation for details.
 468      * Note that this operation may be significantly slower than hasBoundaryBefore().
 469      * @param c character to test
 470      * @return TRUE if c has a normalization boundary after it
 471      * @stable ICU 4.4
 472      */
 473     virtual UBool hasBoundaryAfter(UChar32 c) const = 0;
 474
 475     /**
 476      * Tests if the character is normalization-inert.
 477      * If true, then the character does not change, nor normalization-interact with
 478      * preceding or following characters.
 479      * In other words, a string containing this character can be normalized
 480      * by processing portions before this character and after this
 481      * character independently.
 482      * This is used for iterative normalization. See the class documentation for details.
 483      * Note that this operation may be significantly slower than hasBoundaryBefore().
 484      * @param c character to test
 485      * @return TRUE if c is normalization-inert
 486      * @stable ICU 4.4
 487      */
 488     virtual UBool isInert(UChar32 c) const = 0;
 489 };
 490
 491 /**
 492  * Normalization filtered by a UnicodeSet.
 493  * Normalizes portions of the text contained in the filter set and leaves
 494  * portions not contained in the filter set unchanged.
 495  * Filtering is done via UnicodeSet::span(..., USET_SPAN_SIMPLE).
 496  * Not-in-the-filter text is treated as "is normalized" and "quick check yes".
 497  * This class implements all of (and only) the Normalizer2 API.
 498  * An instance of this class is unmodifiable/immutable but is constructed and
 499  * must be destructed by the owner.
 500  * @stable ICU 4.4
 501  */
 502 class U_COMMON_API FilteredNormalizer2 : public Normalizer2 {
 503 public:
 504     /**
 505      * Constructs a filtered normalizer wrapping any Normalizer2 instance
 506      * and a filter set.
 507      * Both are aliased and must not be modified or deleted while this object
 508      * is used.
 509      * The filter set should be frozen; otherwise the performance will suffer greatly.
 510      * @param n2 wrapped Normalizer2 instance
 511      * @param filterSet UnicodeSet which determines the characters to be normalized
 512      * @stable ICU 4.4
 513      */
 514     FilteredNormalizer2(const Normalizer2 &n2, const UnicodeSet &filterSet) :
 515             norm2(n2), set(filterSet) {}
 516
 517     /**
 518      * Destructor.
 519      * @stable ICU 4.4
 520      */
 521     ~FilteredNormalizer2();
 522
 523     /**
 524      * Writes the normalized form of the source string to the destination string
 525      * (replacing its contents) and returns the destination string.
 526      * The source and destination strings must be different objects.
 527      * @param src source string
 528      * @param dest destination string; its contents is replaced with normalized src
 529      * @param errorCode Standard ICU error code. Its input value must
 530      *                  pass the U_SUCCESS() test, or else the function returns
 531      *                  immediately. Check for U_FAILURE() on output or use with
 532      *                  function chaining. (See User Guide for details.)
 533      * @return dest
 534      * @stable ICU 4.4
 535      */
 536     virtual UnicodeString &
 537     normalize(const UnicodeString &src,
 538               UnicodeString &dest,
 539               UErrorCode &errorCode) const U_OVERRIDE;
 540
 541     /**
 542      * Normalizes a UTF-8 string and optionally records how source substrings
 543      * relate to changed and unchanged result substrings.
 544      *
 545      * Currently implemented completely only for "compose" modes,
 546      * such as for NFC, NFKC, and NFKC_Casefold
 547      * (UNORM2_COMPOSE and UNORM2_COMPOSE_CONTIGUOUS).
 548      * Otherwise currently converts to & from UTF-16 and does not support edits.
 549      *
 550      * @param options   Options bit set, usually 0. See U_OMIT_UNCHANGED_TEXT and U_EDITS_NO_RESET.
 551      * @param src       Source UTF-8 string.
 552      * @param sink      A ByteSink to which the normalized UTF-8 result string is written.
 553      *                  sink.Flush() is called at the end.
 554      * @param edits     Records edits for index mapping, working with styled text,
 555      *                  and getting only changes (if any).
 556      *                  The Edits contents is undefined if any error occurs.
 557      *                  This function calls edits->reset() first unless
 558      *                  options includes U_EDITS_NO_RESET. edits can be nullptr.
 559      * @param errorCode Standard ICU error code. Its input value must
 560      *                  pass the U_SUCCESS() test, or else the function returns
 561      *                  immediately. Check for U_FAILURE() on output or use with
 562      *                  function chaining. (See User Guide for details.)
 563      * @draft ICU 60
 564      */
 565     virtual void
 566     normalizeUTF8(uint32_t options, StringPiece src, ByteSink &sink,
 567                   Edits *edits, UErrorCode &errorCode) const U_OVERRIDE;
 568
 569     /**
 570      * Appends the normalized form of the second string to the first string
 571      * (merging them at the boundary) and returns the first string.
 572      * The result is normalized if the first string was normalized.
 573      * The first and second strings must be different objects.
 574      * @param first string, should be normalized
 575      * @param second string, will be normalized
 576      * @param errorCode Standard ICU error code. Its input value must
 577      *                  pass the U_SUCCESS() test, or else the function returns
 578      *                  immediately. Check for U_FAILURE() on output or use with
 579      *                  function chaining. (See User Guide for details.)
 580      * @return first
 581      * @stable ICU 4.4
 582      */
 583     virtual UnicodeString &
 584     normalizeSecondAndAppend(UnicodeString &first,
 585                              const UnicodeString &second,
 586                              UErrorCode &errorCode) const U_OVERRIDE;
 587     /**
 588      * Appends the second string to the first string
 589      * (merging them at the boundary) and returns the first string.
 590      * The result is normalized if both the strings were normalized.
 591      * The first and second strings must be different objects.
 592      * @param first string, should be normalized
 593      * @param second string, should be normalized
 594      * @param errorCode Standard ICU error code. Its input value must
 595      *                  pass the U_SUCCESS() test, or else the function returns
 596      *                  immediately. Check for U_FAILURE() on output or use with
 597      *                  function chaining. (See User Guide for details.)
 598      * @return first
 599      * @stable ICU 4.4
 600      */
 601     virtual UnicodeString &
 602     append(UnicodeString &first,
 603            const UnicodeString &second,
 604            UErrorCode &errorCode) const U_OVERRIDE;
 605
 606     /**
 607      * Gets the decomposition mapping of c.
 608      * For details see the base class documentation.
 609      *
 610      * This function is independent of the mode of the Normalizer2.
 611      * @param c code point
 612      * @param decomposition String object which will be set to c's
 613      *                      decomposition mapping, if there is one.
 614      * @return TRUE if c has a decomposition, otherwise FALSE
 615      * @stable ICU 4.6
 616      */
 617     virtual UBool
 618     getDecomposition(UChar32 c, UnicodeString &decomposition) const U_OVERRIDE;
 619
 620     /**
 621      * Gets the raw decomposition mapping of c.
 622      * For details see the base class documentation.
 623      *
 624      * This function is independent of the mode of the Normalizer2.
 625      * @param c code point
 626      * @param decomposition String object which will be set to c's
 627      *                      raw decomposition mapping, if there is one.
 628      * @return TRUE if c has a decomposition, otherwise FALSE
 629      * @stable ICU 49
 630      */
 631     virtual UBool
 632     getRawDecomposition(UChar32 c, UnicodeString &decomposition) const U_OVERRIDE;
 633
 634     /**
 635      * Performs pairwise composition of a & b and returns the composite if there is one.
 636      * For details see the base class documentation.
 637      *
 638      * This function is independent of the mode of the Normalizer2.
 639      * @param a A (normalization starter) code point.
 640      * @param b Another code point.
 641      * @return The non-negative composite code point if there is one; otherwise a negative value.
 642      * @stable ICU 49
 643      */
 644     virtual UChar32
 645     composePair(UChar32 a, UChar32 b) const U_OVERRIDE;
 646
 647     /**
 648      * Gets the combining class of c.
 649      * The default implementation returns 0
 650      * but all standard implementations return the Unicode Canonical_Combining_Class value.
 651      * @param c code point
 652      * @return c's combining class
 653      * @stable ICU 49
 654      */
 655     virtual uint8_t
 656     getCombiningClass(UChar32 c) const U_OVERRIDE;
 657
 658     /**
 659      * Tests if the string is normalized.
 660      * For details see the Normalizer2 base class documentation.
 661      * @param s input string
 662      * @param errorCode Standard ICU error code. Its input value must
 663      *                  pass the U_SUCCESS() test, or else the function returns
 664      *                  immediately. Check for U_FAILURE() on output or use with
 665      *                  function chaining. (See User Guide for details.)
 666      * @return TRUE if s is normalized
 667      * @stable ICU 4.4
 668      */
 669     virtual UBool
 670     isNormalized(const UnicodeString &s, UErrorCode &errorCode) const U_OVERRIDE;
 671     /**
 672      * Tests if the UTF-8 string is normalized.
 673      * Internally, in cases where the quickCheck() method would return "maybe"
 674      * (which is only possible for the two COMPOSE modes) this method
 675      * resolves to "yes" or "no" to provide a definitive result,
 676      * at the cost of doing more work in those cases.
 677      *
 678      * This works for all normalization modes,
 679      * but it is currently optimized for UTF-8 only for "compose" modes,
 680      * such as for NFC, NFKC, and NFKC_Casefold
 681      * (UNORM2_COMPOSE and UNORM2_COMPOSE_CONTIGUOUS).
 682      * For other modes it currently converts to UTF-16 and calls isNormalized().
 683      *
 684      * @param s UTF-8 input string
 685      * @param errorCode Standard ICU error code. Its input value must
 686      *                  pass the U_SUCCESS() test, or else the function returns
 687      *                  immediately. Check for U_FAILURE() on output or use with
 688      *                  function chaining. (See User Guide for details.)
 689      * @return TRUE if s is normalized
 690      * @draft ICU 60
 691      */
 692     virtual UBool
 693     isNormalizedUTF8(StringPiece s, UErrorCode &errorCode) const U_OVERRIDE;
 694     /**
 695      * Tests if the string is normalized.
 696      * For details see the Normalizer2 base class documentation.
 697      * @param s input string
 698      * @param errorCode Standard ICU error code. Its input value must
 699      *                  pass the U_SUCCESS() test, or else the function returns
 700      *                  immediately. Check for U_FAILURE() on output or use with
 701      *                  function chaining. (See User Guide for details.)
 702      * @return UNormalizationCheckResult
 703      * @stable ICU 4.4
 704      */
 705     virtual UNormalizationCheckResult
 706     quickCheck(const UnicodeString &s, UErrorCode &errorCode) const U_OVERRIDE;
 707     /**
 708      * Returns the end of the normalized substring of the input string.
 709      * For details see the Normalizer2 base class documentation.
 710      * @param s input string
 711      * @param errorCode Standard ICU error code. Its input value must
 712      *                  pass the U_SUCCESS() test, or else the function returns
 713      *                  immediately. Check for U_FAILURE() on output or use with
 714      *                  function chaining. (See User Guide for details.)
 715      * @return "yes" span end index
 716      * @stable ICU 4.4
 717      */
 718     virtual int32_t
 719     spanQuickCheckYes(const UnicodeString &s, UErrorCode &errorCode) const U_OVERRIDE;
 720
 721     /**
 722      * Tests if the character always has a normalization boundary before it,
 723      * regardless of context.
 724      * For details see the Normalizer2 base class documentation.
 725      * @param c character to test
 726      * @return TRUE if c has a normalization boundary before it
 727      * @stable ICU 4.4
 728      */
 729     virtual UBool hasBoundaryBefore(UChar32 c) const U_OVERRIDE;
 730
 731     /**
 732      * Tests if the character always has a normalization boundary after it,
 733      * regardless of context.
 734      * For details see the Normalizer2 base class documentation.
 735      * @param c character to test
 736      * @return TRUE if c has a normalization boundary after it
 737      * @stable ICU 4.4
 738      */
 739     virtual UBool hasBoundaryAfter(UChar32 c) const U_OVERRIDE;
 740
 741     /**
 742      * Tests if the character is normalization-inert.
 743      * For details see the Normalizer2 base class documentation.
 744      * @param c character to test
 745      * @return TRUE if c is normalization-inert
 746      * @stable ICU 4.4
 747      */
 748     virtual UBool isInert(UChar32 c) const U_OVERRIDE;
 749 private:
 750     UnicodeString &
 751     normalize(const UnicodeString &src,
 752               UnicodeString &dest,
 753               USetSpanCondition spanCondition,
 754               UErrorCode &errorCode) const;
 755
 756     void
 757     normalizeUTF8(uint32_t options, const char *src, int32_t length,
 758                   ByteSink &sink, Edits *edits,
 759                   USetSpanCondition spanCondition,
 760                   UErrorCode &errorCode) const;
 761
 762     UnicodeString &
 763     normalizeSecondAndAppend(UnicodeString &first,
 764                              const UnicodeString &second,
 765                              UBool doNormalize,
 766                              UErrorCode &errorCode) const;
 767
 768     const Normalizer2 &norm2;
 769     const UnicodeSet &set;
 770 };
 771
 772 U_NAMESPACE_END
 773 #endif // U_SHOW_CPLUSPLUS_API
 774
 775 #endif  // !UCONFIG_NO_NORMALIZATION
 776 #endif  // __NORMALIZER2_H__