icuSources/common/unicode/brkiter.h

   1 /*
   2 ********************************************************************************
   3 *   Copyright (C) 1997-2007, International Business Machines
   4 *   Corporation and others.  All Rights Reserved.
   5 ********************************************************************************
   6 *
   7 * File brkiter.h
   8 *
   9 * Modification History:
  10 *
  11 *   Date        Name        Description
  12 *   02/18/97    aliu        Added typedef for TextCount.  Made DONE const.
  13 *   05/07/97    aliu        Fixed DLL declaration.
  14 *   07/09/97    jfitz       Renamed BreakIterator and interface synced with JDK
  15 *   08/11/98    helena      Sync-up JDK1.2.
  16 *   01/13/2000  helena      Added UErrorCode parameter to createXXXInstance methods.
  17 ********************************************************************************
  18 */
  19
  20 #ifndef BRKITER_H
  21 #define BRKITER_H
  22
  23 #include "unicode/utypes.h"
  24
  25 /**
  26  * \file
  27  * \brief C++ API: Break Iterator.
  28  */
  29
  30 #if UCONFIG_NO_BREAK_ITERATION
  31
  32 U_NAMESPACE_BEGIN
  33
  34 /*
  35  * Allow the declaration of APIs with pointers to BreakIterator
  36  * even when break iteration is removed from the build.
  37  */
  38 class BreakIterator;
  39
  40 U_NAMESPACE_END
  41
  42 #else
  43
  44 #include "unicode/uobject.h"
  45 #include "unicode/unistr.h"
  46 #include "unicode/chariter.h"
  47 #include "unicode/locid.h"
  48 #include "unicode/ubrk.h"
  49 #include "unicode/strenum.h"
  50 #include "unicode/utext.h"
  51 #include "unicode/umisc.h"
  52
  53 U_NAMESPACE_BEGIN
  54
  55 /**
  56  * The BreakIterator class implements methods for finding the location
  57  * of boundaries in text. BreakIterator is an abstract base class.
  58  * Instances of BreakIterator maintain a current position and scan over
  59  * text returning the index of characters where boundaries occur.
  60  * <p>
  61  * Line boundary analysis determines where a text string can be broken
  62  * when line-wrapping. The mechanism correctly handles punctuation and
  63  * hyphenated words.
  64  * <p>
  65  * Sentence boundary analysis allows selection with correct
  66  * interpretation of periods within numbers and abbreviations, and
  67  * trailing punctuation marks such as quotation marks and parentheses.
  68  * <p>
  69  * Word boundary analysis is used by search and replace functions, as
  70  * well as within text editing applications that allow the user to
  71  * select words with a double click. Word selection provides correct
  72  * interpretation of punctuation marks within and following
  73  * words. Characters that are not part of a word, such as symbols or
  74  * punctuation marks, have word-breaks on both sides.
  75  * <p>
  76  * Character boundary analysis allows users to interact with
  77  * characters as they expect to, for example, when moving the cursor
  78  * through a text string. Character boundary analysis provides correct
  79  * navigation of through character strings, regardless of how the
  80  * character is stored.  For example, an accented character might be
  81  * stored as a base character and a diacritical mark. What users
  82  * consider to be a character can differ between languages.
  83  * <p>
  84  * The text boundary positions are found according to the rules
  85  * described in Unicode Standard Annex #29, Text Boundaries, and
  86  * Unicode Standard Annex #14, Line Breaking Properties.  These
  87  * are available at http://www.unicode.org/reports/tr14/ and
  88  * http://www.unicode.org/reports/tr29/.
  89  * <p>
  90  * In addition to the C++ API defined in this header file, a
  91  * plain C API with equivalent functionality is defined in the
  92  * file ubrk.h
  93  * <p>
  94  * Code snippits illustrating the use of the Break Iterator APIs
  95  * are available in the ICU User Guide,
  96  * http://icu-project.org/userguide/boundaryAnalysis.html
  97  * and in the sample program icu/source/samples/break/break.cpp"
  98  *
  99  */
 100 class U_COMMON_API BreakIterator : public UObject {
 101 public:
 102     /**
 103      *  destructor
 104      *  @stable ICU 2.0
 105      */
 106     virtual ~BreakIterator();
 107
 108     /**
 109      * Return true if another object is semantically equal to this
 110      * one. The other object should be an instance of the same subclass of
 111      * BreakIterator. Objects of different subclasses are considered
 112      * unequal.
 113      * <P>
 114      * Return true if this BreakIterator is at the same position in the
 115      * same text, and is the same class and type (word, line, etc.) of
 116      * BreakIterator, as the argument.  Text is considered the same if
 117      * it contains the same characters, it need not be the same
 118      * object, and styles are not considered.
 119      * @stable ICU 2.0
 120      */
 121     virtual UBool operator==(const BreakIterator&) const = 0;
 122
 123     /**
 124      * Returns the complement of the result of operator==
 125      * @param rhs The BreakIterator to be compared for inequality
 126      * @return the complement of the result of operator==
 127      * @stable ICU 2.0
 128      */
 129     UBool operator!=(const BreakIterator& rhs) const { return !operator==(rhs); }
 130
 131     /**
 132      * Return a polymorphic copy of this object.  This is an abstract
 133      * method which subclasses implement.
 134      * @stable ICU 2.0
 135      */
 136     virtual BreakIterator* clone(void) const = 0;
 137
 138     /**
 139      * Return a polymorphic class ID for this object. Different subclasses
 140      * will return distinct unequal values.
 141      * @stable ICU 2.0
 142      */
 143     virtual UClassID getDynamicClassID(void) const = 0;
 144
 145     /**
 146      * Return a CharacterIterator over the text being analyzed.
 147      * @stable ICU 2.0
 148      */
 149     virtual CharacterIterator& getText(void) const = 0;
 150
 151
 152     /**
 153       *  Get a UText for the text being analyzed.
 154       *  The returned UText is a shallow clone of the UText used internally
 155       *  by the break iterator implementation.  It can safely be used to
 156       *  access the text without impacting any break iterator operations,
 157       *  but the underlying text itself must not be altered.
 158       *
 159       * @param fillIn A UText to be filled in.  If NULL, a new UText will be
 160       *           allocated to hold the result.
 161       * @param status receives any error codes.
 162       * @return   The current UText for this break iterator.  If an input
 163       *           UText was provided, it will always be returned.
 164       * @stable ICU 3.4
 165       */
 166      virtual UText *getUText(UText *fillIn, UErrorCode &status) const = 0;
 167
 168     /**
 169      * Change the text over which this operates. The text boundary is
 170      * reset to the start.
 171      * @param text The UnicodeString used to change the text.
 172      * @stable ICU 2.0
 173      */
 174     virtual void  setText(const UnicodeString &text) = 0;
 175
 176     /**
 177      * Reset the break iterator to operate over the text represented by
 178      * the UText.  The iterator position is reset to the start.
 179      *
 180      * This function makes a shallow clone of the supplied UText.  This means
 181      * that the caller is free to immediately close or otherwise reuse the
 182      * Utext that was passed as a parameter, but that the underlying text itself
 183      * must not be altered while being referenced by the break iterator.
 184      *
 185      * @param text The UText used to change the text.
 186      * @param status receives any error codes.
 187      * @stable ICU 3.4
 188      */
 189     virtual void  setText(UText *text, UErrorCode &status) = 0;
 190
 191     /**
 192      * Change the text over which this operates. The text boundary is
 193      * reset to the start.
 194      * Note that setText(UText *) provides similar functionality to this function,
 195      * and is more efficient.
 196      * @param it The CharacterIterator used to change the text.
 197      * @stable ICU 2.0
 198      */
 199     virtual void  adoptText(CharacterIterator* it) = 0;
 200
 201     enum {
 202         /**
 203          * DONE is returned by previous() and next() after all valid
 204          * boundaries have been returned.
 205          * @stable ICU 2.0
 206          */
 207         DONE = (int32_t)-1
 208     };
 209
 210     /**
 211      * Return the index of the first character in the text being scanned.
 212      * @stable ICU 2.0
 213      */
 214     virtual int32_t first(void) = 0;
 215
 216     /**
 217      * Return the index immediately BEYOND the last character in the text being scanned.
 218      * @stable ICU 2.0
 219      */
 220     virtual int32_t last(void) = 0;
 221
 222     /**
 223      * Return the boundary preceding the current boundary.
 224      * @return The character index of the previous text boundary or DONE if all
 225      * boundaries have been returned.
 226      * @stable ICU 2.0
 227      */
 228     virtual int32_t previous(void) = 0;
 229
 230     /**
 231      * Return the boundary following the current boundary.
 232      * @return The character index of the next text boundary or DONE if all
 233      * boundaries have been returned.
 234      * @stable ICU 2.0
 235      */
 236     virtual int32_t next(void) = 0;
 237
 238     /**
 239      * Return character index of the current interator position within the text.
 240      * @return The boundary most recently returned.
 241      * @stable ICU 2.0
 242      */
 243     virtual int32_t current(void) const = 0;
 244
 245     /**
 246      * Return the first boundary following the specified offset.
 247      * The value returned is always greater than the offset or
 248      * the value BreakIterator.DONE
 249      * @param offset the offset to begin scanning.
 250      * @return The first boundary after the specified offset.
 251      * @stable ICU 2.0
 252      */
 253     virtual int32_t following(int32_t offset) = 0;
 254
 255     /**
 256      * Return the first boundary preceding the specified offset.
 257      * The value returned is always smaller than the offset or
 258      * the value BreakIterator.DONE
 259      * @param offset the offset to begin scanning.
 260      * @return The first boundary before the specified offset.
 261      * @stable ICU 2.0
 262      */
 263     virtual int32_t preceding(int32_t offset) = 0;
 264
 265     /**
 266      * Return true if the specfied position is a boundary position.
 267      * As a side effect, the current position of the iterator is set
 268      * to the first boundary position at or following the specified offset.
 269      * @param offset the offset to check.
 270      * @return True if "offset" is a boundary position.
 271      * @stable ICU 2.0
 272      */
 273     virtual UBool isBoundary(int32_t offset) = 0;
 274
 275     /**
 276      * Return the nth boundary from the current boundary
 277      * @param n which boundary to return.  A value of 0
 278      * does nothing.  Negative values move to previous boundaries
 279      * and positive values move to later boundaries.
 280      * @return The index of the nth boundary from the current position, or
 281      * DONE if there are fewer than |n| boundaries in the specfied direction.
 282      * @stable ICU 2.0
 283      */
 284     virtual int32_t next(int32_t n) = 0;
 285
 286     /**
 287      * Create BreakIterator for word-breaks using the given locale.
 288      * Returns an instance of a BreakIterator implementing word breaks.
 289      * WordBreak is useful for word selection (ex. double click)
 290      * @param where the locale.
 291      * @param status the error code
 292      * @return A BreakIterator for word-breaks.  The UErrorCode& status
 293      * parameter is used to return status information to the user.
 294      * To check whether the construction succeeded or not, you should check
 295      * the value of U_SUCCESS(err).  If you wish more detailed information, you
 296      * can check for informational error results which still indicate success.
 297      * U_USING_FALLBACK_WARNING indicates that a fall back locale was used.  For
 298      * example, 'de_CH' was requested, but nothing was found there, so 'de' was
 299      * used.  U_USING_DEFAULT_WARNING indicates that the default locale data was
 300      * used; neither the requested locale nor any of its fall back locales
 301      * could be found.
 302      * The caller owns the returned object and is responsible for deleting it.
 303      * @stable ICU 2.0
 304      */
 305     static BreakIterator* U_EXPORT2
 306     createWordInstance(const Locale& where, UErrorCode& status);
 307
 308     /**
 309      * Create BreakIterator for line-breaks using specified locale.
 310      * Returns an instance of a BreakIterator implementing line breaks. Line
 311      * breaks are logically possible line breaks, actual line breaks are
 312      * usually determined based on display width.
 313      * LineBreak is useful for word wrapping text.
 314      * @param where the locale.
 315      * @param status The error code.
 316      * @return A BreakIterator for line-breaks.  The UErrorCode& status
 317      * parameter is used to return status information to the user.
 318      * To check whether the construction succeeded or not, you should check
 319      * the value of U_SUCCESS(err).  If you wish more detailed information, you
 320      * can check for informational error results which still indicate success.
 321      * U_USING_FALLBACK_WARNING indicates that a fall back locale was used.  For
 322      * example, 'de_CH' was requested, but nothing was found there, so 'de' was
 323      * used.  U_USING_DEFAULT_WARNING indicates that the default locale data was
 324      * used; neither the requested locale nor any of its fall back locales
 325      * could be found.
 326      * The caller owns the returned object and is responsible for deleting it.
 327      * @stable ICU 2.0
 328      */
 329     static BreakIterator* U_EXPORT2
 330     createLineInstance(const Locale& where, UErrorCode& status);
 331
 332     /**
 333      * Create BreakIterator for character-breaks using specified locale
 334      * Returns an instance of a BreakIterator implementing character breaks.
 335      * Character breaks are boundaries of combining character sequences.
 336      * @param where the locale.
 337      * @param status The error code.
 338      * @return A BreakIterator for character-breaks.  The UErrorCode& status
 339      * parameter is used to return status information to the user.
 340      * To check whether the construction succeeded or not, you should check
 341      * the value of U_SUCCESS(err).  If you wish more detailed information, you
 342      * can check for informational error results which still indicate success.
 343      * U_USING_FALLBACK_WARNING indicates that a fall back locale was used.  For
 344      * example, 'de_CH' was requested, but nothing was found there, so 'de' was
 345      * used.  U_USING_DEFAULT_WARNING indicates that the default locale data was
 346      * used; neither the requested locale nor any of its fall back locales
 347      * could be found.
 348      * The caller owns the returned object and is responsible for deleting it.
 349      * @stable ICU 2.0
 350      */
 351     static BreakIterator* U_EXPORT2
 352     createCharacterInstance(const Locale& where, UErrorCode& status);
 353
 354     /**
 355      * Create BreakIterator for sentence-breaks using specified locale
 356      * Returns an instance of a BreakIterator implementing sentence breaks.
 357      * @param where the locale.
 358      * @param status The error code.
 359      * @return A BreakIterator for sentence-breaks.  The UErrorCode& status
 360      * parameter is used to return status information to the user.
 361      * To check whether the construction succeeded or not, you should check
 362      * the value of U_SUCCESS(err).  If you wish more detailed information, you
 363      * can check for informational error results which still indicate success.
 364      * U_USING_FALLBACK_WARNING indicates that a fall back locale was used.  For
 365      * example, 'de_CH' was requested, but nothing was found there, so 'de' was
 366      * used.  U_USING_DEFAULT_WARNING indicates that the default locale data was
 367      * used; neither the requested locale nor any of its fall back locales
 368      * could be found.
 369      * The caller owns the returned object and is responsible for deleting it.
 370      * @stable ICU 2.0
 371      */
 372     static BreakIterator* U_EXPORT2
 373     createSentenceInstance(const Locale& where, UErrorCode& status);
 374
 375     /**
 376      * Create BreakIterator for title-casing breaks using the specified locale
 377      * Returns an instance of a BreakIterator implementing title breaks.
 378      * The iterator returned locates title boundaries as described for
 379      * Unicode 3.2 only. For Unicode 4.0 and above title boundary iteration,
 380      * please use Word Boundary iterator.{@link #createWordInstance }
 381      *
 382      * @param where the locale.
 383      * @param status The error code.
 384      * @return A BreakIterator for title-breaks.  The UErrorCode& status
 385      * parameter is used to return status information to the user.
 386      * To check whether the construction succeeded or not, you should check
 387      * the value of U_SUCCESS(err).  If you wish more detailed information, you
 388      * can check for informational error results which still indicate success.
 389      * U_USING_FALLBACK_WARNING indicates that a fall back locale was used.  For
 390      * example, 'de_CH' was requested, but nothing was found there, so 'de' was
 391      * used.  U_USING_DEFAULT_WARNING indicates that the default locale data was
 392      * used; neither the requested locale nor any of its fall back locales
 393      * could be found.
 394      * The caller owns the returned object and is responsible for deleting it.
 395      * @stable ICU 2.1
 396      */
 397     static BreakIterator* U_EXPORT2
 398     createTitleInstance(const Locale& where, UErrorCode& status);
 399
 400     /**
 401      * Get the set of Locales for which TextBoundaries are installed.
 402      * <p><b>Note:</b> this will not return locales added through the register
 403      * call. To see the registered locales too, use the getAvailableLocales
 404      * function that returns a StringEnumeration object </p>
 405      * @param count the output parameter of number of elements in the locale list
 406      * @return available locales
 407      * @stable ICU 2.0
 408      */
 409     static const Locale* U_EXPORT2 getAvailableLocales(int32_t& count);
 410
 411     /**
 412      * Get name of the object for the desired Locale, in the desired langauge.
 413      * @param objectLocale must be from getAvailableLocales.
 414      * @param displayLocale specifies the desired locale for output.
 415      * @param name the fill-in parameter of the return value
 416      * Uses best match.
 417      * @return user-displayable name
 418      * @stable ICU 2.0
 419      */
 420     static UnicodeString& U_EXPORT2 getDisplayName(const Locale& objectLocale,
 421                                          const Locale& displayLocale,
 422                                          UnicodeString& name);
 423
 424     /**
 425      * Get name of the object for the desired Locale, in the langauge of the
 426      * default locale.
 427      * @param objectLocale must be from getMatchingLocales
 428      * @param name the fill-in parameter of the return value
 429      * @return user-displayable name
 430      * @stable ICU 2.0
 431      */
 432     static UnicodeString& U_EXPORT2 getDisplayName(const Locale& objectLocale,
 433                                          UnicodeString& name);
 434
 435     /**
 436      * Thread safe client-buffer-based cloning operation
 437      *    Do NOT call delete on a safeclone, since 'new' is not used to create it.
 438      * @param stackBuffer user allocated space for the new clone. If NULL new memory will be allocated.
 439      * If buffer is not large enough, new memory will be allocated.
 440      * @param BufferSize reference to size of allocated space.
 441      * If BufferSize == 0, a sufficient size for use in cloning will
 442      * be returned ('pre-flighting')
 443      * If BufferSize is not enough for a stack-based safe clone,
 444      * new memory will be allocated.
 445      * @param status to indicate whether the operation went on smoothly or there were errors
 446      *  An informational status value, U_SAFECLONE_ALLOCATED_ERROR, is used if any allocations were
 447      *  necessary.
 448      * @return pointer to the new clone
 449      *
 450      * @stable ICU 2.0
 451      */
 452     virtual BreakIterator *  createBufferClone(void *stackBuffer,
 453                                                int32_t &BufferSize,
 454                                                UErrorCode &status) = 0;
 455
 456     /**
 457      *   Determine whether the BreakIterator was created in user memory by
 458      *   createBufferClone(), and thus should not be deleted.  Such objects
 459      *   must be closed by an explicit call to the destructor (not delete).
 460      *  @stable ICU 2.0
 461      */
 462     inline UBool isBufferClone(void);
 463
 464 #if !UCONFIG_NO_SERVICE
 465     /**
 466      * Register a new break iterator of the indicated kind, to use in the given locale.
 467      * The break iterator will be adopted.  Clones of the iterator will be returned
 468      * if a request for a break iterator of the given kind matches or falls back to
 469      * this locale.
 470      * @param toAdopt the BreakIterator instance to be adopted
 471      * @param locale the Locale for which this instance is to be registered
 472      * @param kind the type of iterator for which this instance is to be registered
 473      * @param status the in/out status code, no special meanings are assigned
 474      * @return a registry key that can be used to unregister this instance
 475      * @stable ICU 2.4
 476      */
 477     static URegistryKey U_EXPORT2 registerInstance(BreakIterator* toAdopt,
 478                                         const Locale& locale,
 479                                         UBreakIteratorType kind,
 480                                         UErrorCode& status);
 481
 482     /**
 483      * Unregister a previously-registered BreakIterator using the key returned from the
 484      * register call.  Key becomes invalid after a successful call and should not be used again.
 485      * The BreakIterator corresponding to the key will be deleted.
 486      * @param key the registry key returned by a previous call to registerInstance
 487      * @param status the in/out status code, no special meanings are assigned
 488      * @return TRUE if the iterator for the key was successfully unregistered
 489      * @stable ICU 2.4
 490      */
 491     static UBool U_EXPORT2 unregister(URegistryKey key, UErrorCode& status);
 492
 493     /**
 494      * Return a StringEnumeration over the locales available at the time of the call,
 495      * including registered locales.
 496      * @return a StringEnumeration over the locales available at the time of the call
 497      * @stable ICU 2.4
 498      */
 499     static StringEnumeration* U_EXPORT2 getAvailableLocales(void);
 500 #endif
 501
 502     /**
 503      * Returns the locale for this break iterator. Two flavors are available: valid and
 504      * actual locale.
 505      * @stable ICU 2.8
 506      */
 507     Locale getLocale(ULocDataLocaleType type, UErrorCode& status) const;
 508
 509     /** Get the locale for this break iterator object. You can choose between valid and actual locale.
 510      *  @param type type of the locale we're looking for (valid or actual)
 511      *  @param status error code for the operation
 512      *  @return the locale
 513      *  @internal
 514      */
 515     const char *getLocaleID(ULocDataLocaleType type, UErrorCode& status) const;
 516
 517  private:
 518     static BreakIterator* buildInstance(const Locale& loc, const char *type, int32_t kind, UErrorCode& status);
 519     static BreakIterator* createInstance(const Locale& loc, int32_t kind, UErrorCode& status);
 520     static BreakIterator* makeInstance(const Locale& loc, int32_t kind, UErrorCode& status);
 521
 522     friend class ICUBreakIteratorFactory;
 523     friend class ICUBreakIteratorService;
 524
 525 protected:
 526     /** @internal */
 527     BreakIterator();
 528     /** @internal */
 529     UBool fBufferClone;
 530     /** @internal */
 531     BreakIterator (const BreakIterator &other) : UObject(other), fBufferClone(FALSE) {}
 532
 533 private:
 534
 535     /** @internal */
 536     char actualLocale[ULOC_FULLNAME_CAPACITY];
 537     char validLocale[ULOC_FULLNAME_CAPACITY];
 538
 539     /**
 540      * The assignment operator has no real implementation.
 541      * It's provided to make the compiler happy. Do not call.
 542      */
 543     BreakIterator& operator=(const BreakIterator&);
 544 };
 545
 546 inline UBool BreakIterator::isBufferClone()
 547 {
 548     return fBufferClone;
 549 }
 550
 551 U_NAMESPACE_END
 552
 553 #endif /* #if !UCONFIG_NO_BREAK_ITERATION */
 554
 555 #endif // _BRKITER
 556 //eof
 557