icuSources/common/unicode/unorm2.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) 2009-2015, International Business Machines
   7 *   Corporation and others.  All Rights Reserved.
   8 *
   9 *******************************************************************************
  10 *   file name:  unorm2.h
  11 *   encoding:   UTF-8
  12 *   tab size:   8 (not used)
  13 *   indentation:4
  14 *
  15 *   created on: 2009dec15
  16 *   created by: Markus W. Scherer
  17 */
  18
  19 #ifndef __UNORM2_H__
  20 #define __UNORM2_H__
  21
  22 /**
  23  * \file
  24  * \brief C API: New API for Unicode Normalization.
  25  *
  26  * Unicode normalization functionality for standard Unicode normalization or
  27  * for using custom mapping tables.
  28  * All instances of UNormalizer2 are unmodifiable/immutable.
  29  * Instances returned by unorm2_getInstance() are singletons that must not be deleted by the caller.
  30  * For more details see the Normalizer2 C++ class.
  31  */
  32
  33 #include "unicode/utypes.h"
  34 #include "unicode/localpointer.h"
  35 #include "unicode/uset.h"
  36
  37 /**
  38  * Constants for normalization modes.
  39  * For details about standard Unicode normalization forms
  40  * and about the algorithms which are also used with custom mapping tables
  41  * see http://www.unicode.org/unicode/reports/tr15/
  42  * @stable ICU 4.4
  43  */
  44 typedef enum {
  45     /**
  46      * Decomposition followed by composition.
  47      * Same as standard NFC when using an "nfc" instance.
  48      * Same as standard NFKC when using an "nfkc" instance.
  49      * For details about standard Unicode normalization forms
  50      * see http://www.unicode.org/unicode/reports/tr15/
  51      * @stable ICU 4.4
  52      */
  53     UNORM2_COMPOSE,
  54     /**
  55      * Map, and reorder canonically.
  56      * Same as standard NFD when using an "nfc" instance.
  57      * Same as standard NFKD when using an "nfkc" instance.
  58      * For details about standard Unicode normalization forms
  59      * see http://www.unicode.org/unicode/reports/tr15/
  60      * @stable ICU 4.4
  61      */
  62     UNORM2_DECOMPOSE,
  63     /**
  64      * "Fast C or D" form.
  65      * If a string is in this form, then further decomposition <i>without reordering</i>
  66      * would yield the same form as DECOMPOSE.
  67      * Text in "Fast C or D" form can be processed efficiently with data tables
  68      * that are "canonically closed", that is, that provide equivalent data for
  69      * equivalent text, without having to be fully normalized.
  70      * Not a standard Unicode normalization form.
  71      * Not a unique form: Different FCD strings can be canonically equivalent.
  72      * For details see http://www.unicode.org/notes/tn5/#FCD
  73      * @stable ICU 4.4
  74      */
  75     UNORM2_FCD,
  76     /**
  77      * Compose only contiguously.
  78      * Also known as "FCC" or "Fast C Contiguous".
  79      * The result will often but not always be in NFC.
  80      * The result will conform to FCD which is useful for processing.
  81      * Not a standard Unicode normalization form.
  82      * For details see http://www.unicode.org/notes/tn5/#FCC
  83      * @stable ICU 4.4
  84      */
  85     UNORM2_COMPOSE_CONTIGUOUS
  86 } UNormalization2Mode;
  87
  88 /**
  89  * Result values for normalization quick check functions.
  90  * For details see http://www.unicode.org/reports/tr15/#Detecting_Normalization_Forms
  91  * @stable ICU 2.0
  92  */
  93 typedef enum UNormalizationCheckResult {
  94   /**
  95    * The input string is not in the normalization form.
  96    * @stable ICU 2.0
  97    */
  98   UNORM_NO,
  99   /**
 100    * The input string is in the normalization form.
 101    * @stable ICU 2.0
 102    */
 103   UNORM_YES,
 104   /**
 105    * The input string may or may not be in the normalization form.
 106    * This value is only returned for composition forms like NFC and FCC,
 107    * when a backward-combining character is found for which the surrounding text
 108    * would have to be analyzed further.
 109    * @stable ICU 2.0
 110    */
 111   UNORM_MAYBE
 112 } UNormalizationCheckResult;
 113
 114 /**
 115  * Opaque C service object type for the new normalization API.
 116  * @stable ICU 4.4
 117  */
 118 struct UNormalizer2;
 119 typedef struct UNormalizer2 UNormalizer2;  /**< C typedef for struct UNormalizer2. @stable ICU 4.4 */
 120
 121 #if !UCONFIG_NO_NORMALIZATION
 122
 123 /**
 124  * Returns a UNormalizer2 instance for Unicode NFC normalization.
 125  * Same as unorm2_getInstance(NULL, "nfc", UNORM2_COMPOSE, pErrorCode).
 126  * Returns an unmodifiable singleton instance. Do not delete it.
 127  * @param pErrorCode Standard ICU error code. Its input value must
 128  *                  pass the U_SUCCESS() test, or else the function returns
 129  *                  immediately. Check for U_FAILURE() on output or use with
 130  *                  function chaining. (See User Guide for details.)
 131  * @return the requested Normalizer2, if successful
 132  * @stable ICU 49
 133  */
 134 U_STABLE const UNormalizer2 * U_EXPORT2
 135 unorm2_getNFCInstance(UErrorCode *pErrorCode);
 136
 137 /**
 138  * Returns a UNormalizer2 instance for Unicode NFD normalization.
 139  * Same as unorm2_getInstance(NULL, "nfc", UNORM2_DECOMPOSE, pErrorCode).
 140  * Returns an unmodifiable singleton instance. Do not delete it.
 141  * @param pErrorCode Standard ICU error code. Its input value must
 142  *                  pass the U_SUCCESS() test, or else the function returns
 143  *                  immediately. Check for U_FAILURE() on output or use with
 144  *                  function chaining. (See User Guide for details.)
 145  * @return the requested Normalizer2, if successful
 146  * @stable ICU 49
 147  */
 148 U_STABLE const UNormalizer2 * U_EXPORT2
 149 unorm2_getNFDInstance(UErrorCode *pErrorCode);
 150
 151 /**
 152  * Returns a UNormalizer2 instance for Unicode NFKC normalization.
 153  * Same as unorm2_getInstance(NULL, "nfkc", UNORM2_COMPOSE, pErrorCode).
 154  * Returns an unmodifiable singleton instance. Do not delete it.
 155  * @param pErrorCode Standard ICU error code. Its input value must
 156  *                  pass the U_SUCCESS() test, or else the function returns
 157  *                  immediately. Check for U_FAILURE() on output or use with
 158  *                  function chaining. (See User Guide for details.)
 159  * @return the requested Normalizer2, if successful
 160  * @stable ICU 49
 161  */
 162 U_STABLE const UNormalizer2 * U_EXPORT2
 163 unorm2_getNFKCInstance(UErrorCode *pErrorCode);
 164
 165 /**
 166  * Returns a UNormalizer2 instance for Unicode NFKD normalization.
 167  * Same as unorm2_getInstance(NULL, "nfkc", UNORM2_DECOMPOSE, pErrorCode).
 168  * Returns an unmodifiable singleton instance. Do not delete it.
 169  * @param pErrorCode Standard ICU error code. Its input value must
 170  *                  pass the U_SUCCESS() test, or else the function returns
 171  *                  immediately. Check for U_FAILURE() on output or use with
 172  *                  function chaining. (See User Guide for details.)
 173  * @return the requested Normalizer2, if successful
 174  * @stable ICU 49
 175  */
 176 U_STABLE const UNormalizer2 * U_EXPORT2
 177 unorm2_getNFKDInstance(UErrorCode *pErrorCode);
 178
 179 /**
 180  * Returns a UNormalizer2 instance for Unicode NFKC_Casefold normalization.
 181  * Same as unorm2_getInstance(NULL, "nfkc_cf", UNORM2_COMPOSE, pErrorCode).
 182  * Returns an unmodifiable singleton instance. Do not delete it.
 183  * @param pErrorCode Standard ICU error code. Its input value must
 184  *                  pass the U_SUCCESS() test, or else the function returns
 185  *                  immediately. Check for U_FAILURE() on output or use with
 186  *                  function chaining. (See User Guide for details.)
 187  * @return the requested Normalizer2, if successful
 188  * @stable ICU 49
 189  */
 190 U_STABLE const UNormalizer2 * U_EXPORT2
 191 unorm2_getNFKCCasefoldInstance(UErrorCode *pErrorCode);
 192
 193 /**
 194  * Returns a UNormalizer2 instance which uses the specified data file
 195  * (packageName/name similar to ucnv_openPackage() and ures_open()/ResourceBundle)
 196  * and which composes or decomposes text according to the specified mode.
 197  * Returns an unmodifiable singleton instance. Do not delete it.
 198  *
 199  * Use packageName=NULL for data files that are part of ICU's own data.
 200  * Use name="nfc" and UNORM2_COMPOSE/UNORM2_DECOMPOSE for Unicode standard NFC/NFD.
 201  * Use name="nfkc" and UNORM2_COMPOSE/UNORM2_DECOMPOSE for Unicode standard NFKC/NFKD.
 202  * Use name="nfkc_cf" and UNORM2_COMPOSE for Unicode standard NFKC_CF=NFKC_Casefold.
 203  *
 204  * @param packageName NULL for ICU built-in data, otherwise application data package name
 205  * @param name "nfc" or "nfkc" or "nfkc_cf" or name of custom data file
 206  * @param mode normalization mode (compose or decompose etc.)
 207  * @param pErrorCode Standard ICU error code. Its input value must
 208  *                  pass the U_SUCCESS() test, or else the function returns
 209  *                  immediately. Check for U_FAILURE() on output or use with
 210  *                  function chaining. (See User Guide for details.)
 211  * @return the requested UNormalizer2, if successful
 212  * @stable ICU 4.4
 213  */
 214 U_STABLE const UNormalizer2 * U_EXPORT2
 215 unorm2_getInstance(const char *packageName,
 216                    const char *name,
 217                    UNormalization2Mode mode,
 218                    UErrorCode *pErrorCode);
 219
 220 /**
 221  * Constructs a filtered normalizer wrapping any UNormalizer2 instance
 222  * and a filter set.
 223  * Both are aliased and must not be modified or deleted while this object
 224  * is used.
 225  * The filter set should be frozen; otherwise the performance will suffer greatly.
 226  * @param norm2 wrapped UNormalizer2 instance
 227  * @param filterSet USet which determines the characters to be normalized
 228  * @param pErrorCode Standard ICU error code. Its input value must
 229  *                   pass the U_SUCCESS() test, or else the function returns
 230  *                   immediately. Check for U_FAILURE() on output or use with
 231  *                   function chaining. (See User Guide for details.)
 232  * @return the requested UNormalizer2, if successful
 233  * @stable ICU 4.4
 234  */
 235 U_STABLE UNormalizer2 * U_EXPORT2
 236 unorm2_openFiltered(const UNormalizer2 *norm2, const USet *filterSet, UErrorCode *pErrorCode);
 237
 238 /**
 239  * Closes a UNormalizer2 instance from unorm2_openFiltered().
 240  * Do not close instances from unorm2_getInstance()!
 241  * @param norm2 UNormalizer2 instance to be closed
 242  * @stable ICU 4.4
 243  */
 244 U_STABLE void U_EXPORT2
 245 unorm2_close(UNormalizer2 *norm2);
 246
 247 #if U_SHOW_CPLUSPLUS_API
 248
 249 U_NAMESPACE_BEGIN
 250
 251 /**
 252  * \class LocalUNormalizer2Pointer
 253  * "Smart pointer" class, closes a UNormalizer2 via unorm2_close().
 254  * For most methods see the LocalPointerBase base class.
 255  *
 256  * @see LocalPointerBase
 257  * @see LocalPointer
 258  * @stable ICU 4.4
 259  */
 260 U_DEFINE_LOCAL_OPEN_POINTER(LocalUNormalizer2Pointer, UNormalizer2, unorm2_close);
 261
 262 U_NAMESPACE_END
 263
 264 #endif // U_SHOW_CPLUSPLUS_API
 265
 266 /**
 267  * Writes the normalized form of the source string to the destination string
 268  * (replacing its contents) and returns the length of the destination string.
 269  * The source and destination strings must be different buffers.
 270  * @param norm2 UNormalizer2 instance
 271  * @param src source string
 272  * @param length length of the source string, or -1 if NUL-terminated
 273  * @param dest destination string; its contents is replaced with normalized src
 274  * @param capacity number of UChars that can be written to dest
 275  * @param pErrorCode Standard ICU error code. Its input value must
 276  *                   pass the U_SUCCESS() test, or else the function returns
 277  *                   immediately. Check for U_FAILURE() on output or use with
 278  *                   function chaining. (See User Guide for details.)
 279  * @return dest
 280  * @stable ICU 4.4
 281  */
 282 U_STABLE int32_t U_EXPORT2
 283 unorm2_normalize(const UNormalizer2 *norm2,
 284                  const UChar *src, int32_t length,
 285                  UChar *dest, int32_t capacity,
 286                  UErrorCode *pErrorCode);
 287 /**
 288  * Appends the normalized form of the second string to the first string
 289  * (merging them at the boundary) and returns the length of the first string.
 290  * The result is normalized if the first string was normalized.
 291  * The first and second strings must be different buffers.
 292  * @param norm2 UNormalizer2 instance
 293  * @param first string, should be normalized
 294  * @param firstLength length of the first string, or -1 if NUL-terminated
 295  * @param firstCapacity number of UChars that can be written to first
 296  * @param second string, will be normalized
 297  * @param secondLength length of the source string, or -1 if NUL-terminated
 298  * @param pErrorCode Standard ICU error code. Its input value must
 299  *                   pass the U_SUCCESS() test, or else the function returns
 300  *                   immediately. Check for U_FAILURE() on output or use with
 301  *                   function chaining. (See User Guide for details.)
 302  * @return first
 303  * @stable ICU 4.4
 304  */
 305 U_STABLE int32_t U_EXPORT2
 306 unorm2_normalizeSecondAndAppend(const UNormalizer2 *norm2,
 307                                 UChar *first, int32_t firstLength, int32_t firstCapacity,
 308                                 const UChar *second, int32_t secondLength,
 309                                 UErrorCode *pErrorCode);
 310 /**
 311  * Appends the second string to the first string
 312  * (merging them at the boundary) and returns the length of the first string.
 313  * The result is normalized if both the strings were normalized.
 314  * The first and second strings must be different buffers.
 315  * @param norm2 UNormalizer2 instance
 316  * @param first string, should be normalized
 317  * @param firstLength length of the first string, or -1 if NUL-terminated
 318  * @param firstCapacity number of UChars that can be written to first
 319  * @param second string, should be normalized
 320  * @param secondLength length of the source string, or -1 if NUL-terminated
 321  * @param pErrorCode Standard ICU error code. Its input value must
 322  *                   pass the U_SUCCESS() test, or else the function returns
 323  *                   immediately. Check for U_FAILURE() on output or use with
 324  *                   function chaining. (See User Guide for details.)
 325  * @return first
 326  * @stable ICU 4.4
 327  */
 328 U_STABLE int32_t U_EXPORT2
 329 unorm2_append(const UNormalizer2 *norm2,
 330               UChar *first, int32_t firstLength, int32_t firstCapacity,
 331               const UChar *second, int32_t secondLength,
 332               UErrorCode *pErrorCode);
 333
 334 /**
 335  * Gets the decomposition mapping of c.
 336  * Roughly equivalent to normalizing the String form of c
 337  * on a UNORM2_DECOMPOSE UNormalizer2 instance, but much faster, and except that this function
 338  * returns a negative value and does not write a string
 339  * if c does not have a decomposition mapping in this instance's data.
 340  * This function is independent of the mode of the UNormalizer2.
 341  * @param norm2 UNormalizer2 instance
 342  * @param c code point
 343  * @param decomposition String buffer which will be set to c's
 344  *                      decomposition mapping, if there is one.
 345  * @param capacity number of UChars that can be written to decomposition
 346  * @param pErrorCode Standard ICU error code. Its input value must
 347  *                   pass the U_SUCCESS() test, or else the function returns
 348  *                   immediately. Check for U_FAILURE() on output or use with
 349  *                   function chaining. (See User Guide for details.)
 350  * @return the non-negative length of c's decomposition, if there is one; otherwise a negative value
 351  * @stable ICU 4.6
 352  */
 353 U_STABLE int32_t U_EXPORT2
 354 unorm2_getDecomposition(const UNormalizer2 *norm2,
 355                         UChar32 c, UChar *decomposition, int32_t capacity,
 356                         UErrorCode *pErrorCode);
 357
 358 /**
 359  * Gets the raw decomposition mapping of c.
 360  *
 361  * This is similar to the unorm2_getDecomposition() function but returns the
 362  * raw decomposition mapping as specified in UnicodeData.txt or
 363  * (for custom data) in the mapping files processed by the gennorm2 tool.
 364  * By contrast, unorm2_getDecomposition() returns the processed,
 365  * recursively-decomposed version of this mapping.
 366  *
 367  * When used on a standard NFKC Normalizer2 instance,
 368  * unorm2_getRawDecomposition() returns the Unicode Decomposition_Mapping (dm) property.
 369  *
 370  * When used on a standard NFC Normalizer2 instance,
 371  * it returns the Decomposition_Mapping only if the Decomposition_Type (dt) is Canonical (Can);
 372  * in this case, the result contains either one or two code points (=1..4 UChars).
 373  *
 374  * This function is independent of the mode of the UNormalizer2.
 375  * @param norm2 UNormalizer2 instance
 376  * @param c code point
 377  * @param decomposition String buffer which will be set to c's
 378  *                      raw decomposition mapping, if there is one.
 379  * @param capacity number of UChars that can be written to decomposition
 380  * @param pErrorCode Standard ICU error code. Its input value must
 381  *                   pass the U_SUCCESS() test, or else the function returns
 382  *                   immediately. Check for U_FAILURE() on output or use with
 383  *                   function chaining. (See User Guide for details.)
 384  * @return the non-negative length of c's raw decomposition, if there is one; otherwise a negative value
 385  * @stable ICU 49
 386  */
 387 U_STABLE int32_t U_EXPORT2
 388 unorm2_getRawDecomposition(const UNormalizer2 *norm2,
 389                            UChar32 c, UChar *decomposition, int32_t capacity,
 390                            UErrorCode *pErrorCode);
 391
 392 /**
 393  * Performs pairwise composition of a & b and returns the composite if there is one.
 394  *
 395  * Returns a composite code point c only if c has a two-way mapping to a+b.
 396  * In standard Unicode normalization, this means that
 397  * c has a canonical decomposition to a+b
 398  * and c does not have the Full_Composition_Exclusion property.
 399  *
 400  * This function is independent of the mode of the UNormalizer2.
 401  * @param norm2 UNormalizer2 instance
 402  * @param a A (normalization starter) code point.
 403  * @param b Another code point.
 404  * @return The non-negative composite code point if there is one; otherwise a negative value.
 405  * @stable ICU 49
 406  */
 407 U_STABLE UChar32 U_EXPORT2
 408 unorm2_composePair(const UNormalizer2 *norm2, UChar32 a, UChar32 b);
 409
 410 /**
 411  * Gets the combining class of c.
 412  * The default implementation returns 0
 413  * but all standard implementations return the Unicode Canonical_Combining_Class value.
 414  * @param norm2 UNormalizer2 instance
 415  * @param c code point
 416  * @return c's combining class
 417  * @stable ICU 49
 418  */
 419 U_STABLE uint8_t U_EXPORT2
 420 unorm2_getCombiningClass(const UNormalizer2 *norm2, UChar32 c);
 421
 422 /**
 423  * Tests if the string is normalized.
 424  * Internally, in cases where the quickCheck() method would return "maybe"
 425  * (which is only possible for the two COMPOSE modes) this method
 426  * resolves to "yes" or "no" to provide a definitive result,
 427  * at the cost of doing more work in those cases.
 428  * @param norm2 UNormalizer2 instance
 429  * @param s input string
 430  * @param length length of the string, or -1 if NUL-terminated
 431  * @param pErrorCode Standard ICU error code. Its input value must
 432  *                   pass the U_SUCCESS() test, or else the function returns
 433  *                   immediately. Check for U_FAILURE() on output or use with
 434  *                   function chaining. (See User Guide for details.)
 435  * @return TRUE if s is normalized
 436  * @stable ICU 4.4
 437  */
 438 U_STABLE UBool U_EXPORT2
 439 unorm2_isNormalized(const UNormalizer2 *norm2,
 440                     const UChar *s, int32_t length,
 441                     UErrorCode *pErrorCode);
 442
 443 /**
 444  * Tests if the string is normalized.
 445  * For the two COMPOSE modes, the result could be "maybe" in cases that
 446  * would take a little more work to resolve definitively.
 447  * Use spanQuickCheckYes() and normalizeSecondAndAppend() for a faster
 448  * combination of quick check + normalization, to avoid
 449  * re-checking the "yes" prefix.
 450  * @param norm2 UNormalizer2 instance
 451  * @param s input string
 452  * @param length length of the string, or -1 if NUL-terminated
 453  * @param pErrorCode Standard ICU error code. Its input value must
 454  *                   pass the U_SUCCESS() test, or else the function returns
 455  *                   immediately. Check for U_FAILURE() on output or use with
 456  *                   function chaining. (See User Guide for details.)
 457  * @return UNormalizationCheckResult
 458  * @stable ICU 4.4
 459  */
 460 U_STABLE UNormalizationCheckResult U_EXPORT2
 461 unorm2_quickCheck(const UNormalizer2 *norm2,
 462                   const UChar *s, int32_t length,
 463                   UErrorCode *pErrorCode);
 464
 465 /**
 466  * Returns the end of the normalized substring of the input string.
 467  * In other words, with <code>end=spanQuickCheckYes(s, ec);</code>
 468  * the substring <code>UnicodeString(s, 0, end)</code>
 469  * will pass the quick check with a "yes" result.
 470  *
 471  * The returned end index is usually one or more characters before the
 472  * "no" or "maybe" character: The end index is at a normalization boundary.
 473  * (See the class documentation for more about normalization boundaries.)
 474  *
 475  * When the goal is a normalized string and most input strings are expected
 476  * to be normalized already, then call this method,
 477  * and if it returns a prefix shorter than the input string,
 478  * copy that prefix and use normalizeSecondAndAppend() for the remainder.
 479  * @param norm2 UNormalizer2 instance
 480  * @param s input string
 481  * @param length length of the string, or -1 if NUL-terminated
 482  * @param pErrorCode Standard ICU error code. Its input value must
 483  *                   pass the U_SUCCESS() test, or else the function returns
 484  *                   immediately. Check for U_FAILURE() on output or use with
 485  *                   function chaining. (See User Guide for details.)
 486  * @return "yes" span end index
 487  * @stable ICU 4.4
 488  */
 489 U_STABLE int32_t U_EXPORT2
 490 unorm2_spanQuickCheckYes(const UNormalizer2 *norm2,
 491                          const UChar *s, int32_t length,
 492                          UErrorCode *pErrorCode);
 493
 494 /**
 495  * Tests if the character always has a normalization boundary before it,
 496  * regardless of context.
 497  * For details see the Normalizer2 base class documentation.
 498  * @param norm2 UNormalizer2 instance
 499  * @param c character to test
 500  * @return TRUE if c has a normalization boundary before it
 501  * @stable ICU 4.4
 502  */
 503 U_STABLE UBool U_EXPORT2
 504 unorm2_hasBoundaryBefore(const UNormalizer2 *norm2, UChar32 c);
 505
 506 /**
 507  * Tests if the character always has a normalization boundary after it,
 508  * regardless of context.
 509  * For details see the Normalizer2 base class documentation.
 510  * @param norm2 UNormalizer2 instance
 511  * @param c character to test
 512  * @return TRUE if c has a normalization boundary after it
 513  * @stable ICU 4.4
 514  */
 515 U_STABLE UBool U_EXPORT2
 516 unorm2_hasBoundaryAfter(const UNormalizer2 *norm2, UChar32 c);
 517
 518 /**
 519  * Tests if the character is normalization-inert.
 520  * For details see the Normalizer2 base class documentation.
 521  * @param norm2 UNormalizer2 instance
 522  * @param c character to test
 523  * @return TRUE if c is normalization-inert
 524  * @stable ICU 4.4
 525  */
 526 U_STABLE UBool U_EXPORT2
 527 unorm2_isInert(const UNormalizer2 *norm2, UChar32 c);
 528
 529 /**
 530  * Option bit for unorm_compare:
 531  * Both input strings are assumed to fulfill FCD conditions.
 532  * @stable ICU 2.2
 533  */
 534 #define UNORM_INPUT_IS_FCD          0x20000
 535
 536 /**
 537  * Option bit for unorm_compare:
 538  * Perform case-insensitive comparison.
 539  * @stable ICU 2.2
 540  */
 541 #define U_COMPARE_IGNORE_CASE       0x10000
 542
 543 #ifndef U_COMPARE_CODE_POINT_ORDER
 544 /* see also unistr.h and ustring.h */
 545 /**
 546  * Option bit for u_strCaseCompare, u_strcasecmp, unorm_compare, etc:
 547  * Compare strings in code point order instead of code unit order.
 548  * @stable ICU 2.2
 549  */
 550 #define U_COMPARE_CODE_POINT_ORDER  0x8000
 551 #endif
 552
 553 /**
 554  * Compares two strings for canonical equivalence.
 555  * Further options include case-insensitive comparison and
 556  * code point order (as opposed to code unit order).
 557  *
 558  * Canonical equivalence between two strings is defined as their normalized
 559  * forms (NFD or NFC) being identical.
 560  * This function compares strings incrementally instead of normalizing
 561  * (and optionally case-folding) both strings entirely,
 562  * improving performance significantly.
 563  *
 564  * Bulk normalization is only necessary if the strings do not fulfill the FCD
 565  * conditions. Only in this case, and only if the strings are relatively long,
 566  * is memory allocated temporarily.
 567  * For FCD strings and short non-FCD strings there is no memory allocation.
 568  *
 569  * Semantically, this is equivalent to
 570  *   strcmp[CodePointOrder](NFD(foldCase(NFD(s1))), NFD(foldCase(NFD(s2))))
 571  * where code point order and foldCase are all optional.
 572  *
 573  * UAX 21 2.5 Caseless Matching specifies that for a canonical caseless match
 574  * the case folding must be performed first, then the normalization.
 575  *
 576  * @param s1 First source string.
 577  * @param length1 Length of first source string, or -1 if NUL-terminated.
 578  *
 579  * @param s2 Second source string.
 580  * @param length2 Length of second source string, or -1 if NUL-terminated.
 581  *
 582  * @param options A bit set of options:
 583  *   - U_FOLD_CASE_DEFAULT or 0 is used for default options:
 584  *     Case-sensitive comparison in code unit order, and the input strings
 585  *     are quick-checked for FCD.
 586  *
 587  *   - UNORM_INPUT_IS_FCD
 588  *     Set if the caller knows that both s1 and s2 fulfill the FCD conditions.
 589  *     If not set, the function will quickCheck for FCD
 590  *     and normalize if necessary.
 591  *
 592  *   - U_COMPARE_CODE_POINT_ORDER
 593  *     Set to choose code point order instead of code unit order
 594  *     (see u_strCompare for details).
 595  *
 596  *   - U_COMPARE_IGNORE_CASE
 597  *     Set to compare strings case-insensitively using case folding,
 598  *     instead of case-sensitively.
 599  *     If set, then the following case folding options are used.
 600  *
 601  *   - Options as used with case-insensitive comparisons, currently:
 602  *
 603  *   - U_FOLD_CASE_EXCLUDE_SPECIAL_I
 604  *    (see u_strCaseCompare for details)
 605  *
 606  *   - regular normalization options shifted left by UNORM_COMPARE_NORM_OPTIONS_SHIFT
 607  *
 608  * @param pErrorCode ICU error code in/out parameter.
 609  *                   Must fulfill U_SUCCESS before the function call.
 610  * @return <0 or 0 or >0 as usual for string comparisons
 611  *
 612  * @see unorm_normalize
 613  * @see UNORM_FCD
 614  * @see u_strCompare
 615  * @see u_strCaseCompare
 616  *
 617  * @stable ICU 2.2
 618  */
 619 U_STABLE int32_t U_EXPORT2
 620 unorm_compare(const UChar *s1, int32_t length1,
 621               const UChar *s2, int32_t length2,
 622               uint32_t options,
 623               UErrorCode *pErrorCode);
 624
 625 #endif  /* !UCONFIG_NO_NORMALIZATION */
 626 #endif  /* __UNORM2_H__ */