ICU 74.1 74.1
brkiter.h
Go to the documentation of this file.
1// © 2016 and later: Unicode, Inc. and others.
2// License & terms of use: http://www.unicode.org/copyright.html
3/*
4********************************************************************************
5* Copyright (C) 1997-2016, International Business Machines
6* Corporation and others. All Rights Reserved.
7********************************************************************************
8*
9* File brkiter.h
10*
11* Modification History:
12*
13* Date Name Description
14* 02/18/97 aliu Added typedef for TextCount. Made DONE const.
15* 05/07/97 aliu Fixed DLL declaration.
16* 07/09/97 jfitz Renamed BreakIterator and interface synced with JDK
17* 08/11/98 helena Sync-up JDK1.2.
18* 01/13/2000 helena Added UErrorCode parameter to createXXXInstance methods.
19********************************************************************************
20*/
21
22#ifndef BRKITER_H
23#define BRKITER_H
24
25#include "unicode/utypes.h"
26
32#include "unicode/utypes.h"
33
34#if U_SHOW_CPLUSPLUS_API
35
36#if UCONFIG_NO_BREAK_ITERATION
37
38U_NAMESPACE_BEGIN
39
40/*
41 * Allow the declaration of APIs with pointers to BreakIterator
42 * even when break iteration is removed from the build.
43 */
44class BreakIterator;
45
46U_NAMESPACE_END
47
48#else
49
50#include "unicode/uobject.h"
51#include "unicode/unistr.h"
52#include "unicode/chariter.h"
53#include "unicode/locid.h"
54#include "unicode/ubrk.h"
55#include "unicode/strenum.h"
56#include "unicode/utext.h"
57#include "unicode/umisc.h"
58
59U_NAMESPACE_BEGIN
60
107public:
112 virtual ~BreakIterator();
113
127 virtual bool operator==(const BreakIterator&) const = 0;
128
135 bool operator!=(const BreakIterator& rhs) const { return !operator==(rhs); }
136
142 virtual BreakIterator* clone() const = 0;
143
149 virtual UClassID getDynamicClassID(void) const override = 0;
150
155 virtual CharacterIterator& getText(void) const = 0;
156
157
172 virtual UText *getUText(UText *fillIn, UErrorCode &status) const = 0;
173
185 virtual void setText(const UnicodeString &text) = 0;
186
205 virtual void setText(UText *text, UErrorCode &status) = 0;
206
215 virtual void adoptText(CharacterIterator* it) = 0;
216
217 enum {
223 DONE = (int32_t)-1
224 };
225
231 virtual int32_t first(void) = 0;
232
238 virtual int32_t last(void) = 0;
239
246 virtual int32_t previous(void) = 0;
247
254 virtual int32_t next(void) = 0;
255
261 virtual int32_t current(void) const = 0;
262
271 virtual int32_t following(int32_t offset) = 0;
272
281 virtual int32_t preceding(int32_t offset) = 0;
282
291 virtual UBool isBoundary(int32_t offset) = 0;
292
302 virtual int32_t next(int32_t n) = 0;
303
317 virtual int32_t getRuleStatus() const;
318
347 virtual int32_t getRuleStatusVec(int32_t *fillInVec, int32_t capacity, UErrorCode &status);
348
368 static BreakIterator* U_EXPORT2
369 createWordInstance(const Locale& where, UErrorCode& status);
370
392 static BreakIterator* U_EXPORT2
393 createLineInstance(const Locale& where, UErrorCode& status);
394
414 static BreakIterator* U_EXPORT2
416
435 static BreakIterator* U_EXPORT2
437
438#ifndef U_HIDE_DEPRECATED_API
461 static BreakIterator* U_EXPORT2
462 createTitleInstance(const Locale& where, UErrorCode& status);
463#endif /* U_HIDE_DEPRECATED_API */
464
474 static const Locale* U_EXPORT2 getAvailableLocales(int32_t& count);
475
485 static UnicodeString& U_EXPORT2 getDisplayName(const Locale& objectLocale,
486 const Locale& displayLocale,
487 UnicodeString& name);
488
497 static UnicodeString& U_EXPORT2 getDisplayName(const Locale& objectLocale,
498 UnicodeString& name);
499
500#ifndef U_FORCE_HIDE_DEPRECATED_API
520 virtual BreakIterator * createBufferClone(void *stackBuffer,
521 int32_t &BufferSize,
522 UErrorCode &status) = 0;
523#endif // U_FORCE_HIDE_DEPRECATED_API
524
525#ifndef U_HIDE_DEPRECATED_API
526
533 inline UBool isBufferClone(void);
534
535#endif /* U_HIDE_DEPRECATED_API */
536
537#if !UCONFIG_NO_SERVICE
553 static URegistryKey U_EXPORT2 registerInstance(BreakIterator* toAdopt,
554 const Locale& locale,
556 UErrorCode& status);
557
570 static UBool U_EXPORT2 unregister(URegistryKey key, UErrorCode& status);
571
578 static StringEnumeration* U_EXPORT2 getAvailableLocales(void);
579#endif
580
587
588#ifndef U_HIDE_INTERNAL_API
595 const char *getLocaleID(ULocDataLocaleType type, UErrorCode& status) const;
596#endif /* U_HIDE_INTERNAL_API */
597
623 virtual BreakIterator &refreshInputText(UText *input, UErrorCode &status) = 0;
624
625 private:
626 static BreakIterator* buildInstance(const Locale& loc, const char *type, UErrorCode& status);
627 static BreakIterator* createInstance(const Locale& loc, int32_t kind, UErrorCode& status);
628 static BreakIterator* makeInstance(const Locale& loc, int32_t kind, UErrorCode& status);
629
630 friend class ICUBreakIteratorFactory;
631 friend class ICUBreakIteratorService;
632
633protected:
634 // Do not enclose protected default/copy constructors with #ifndef U_HIDE_INTERNAL_API
635 // or else the compiler will create a public ones.
640#ifndef U_HIDE_INTERNAL_API
642 BreakIterator (const Locale& valid, const Locale &actual);
644 BreakIterator &operator = (const BreakIterator &other);
645#endif /* U_HIDE_INTERNAL_API */
646
647private:
648
650 char actualLocale[ULOC_FULLNAME_CAPACITY];
651 char validLocale[ULOC_FULLNAME_CAPACITY];
652 char requestLocale[ULOC_FULLNAME_CAPACITY];
653};
654
655#ifndef U_HIDE_DEPRECATED_API
656
657inline UBool BreakIterator::isBufferClone()
658{
659 return false;
660}
661
662#endif /* U_HIDE_DEPRECATED_API */
663
664U_NAMESPACE_END
665
666#endif /* #if !UCONFIG_NO_BREAK_ITERATION */
667
668#endif /* U_SHOW_CPLUSPLUS_API */
669
670#endif // BRKITER_H
671//eof
C++ API: Character Iterator.
The BreakIterator class implements methods for finding the location of boundaries in text.
Definition: brkiter.h:106
virtual bool operator==(const BreakIterator &) const =0
Return true if another object is semantically equal to this one.
virtual ~BreakIterator()
destructor
BreakIterator(const Locale &valid, const Locale &actual)
BreakIterator(const BreakIterator &other)
virtual CharacterIterator & getText(void) const =0
Return a CharacterIterator over the text being analyzed.
virtual int32_t next(void)=0
Advance the iterator to the boundary following the current boundary.
virtual int32_t getRuleStatus() const
For RuleBasedBreakIterators, return the status tag from the break rule that determined the boundary a...
virtual void adoptText(CharacterIterator *it)=0
Change the text over which this operates.
static BreakIterator * createCharacterInstance(const Locale &where, UErrorCode &status)
Create BreakIterator for character-breaks using specified locale Returns an instance of a BreakIterat...
virtual int32_t last(void)=0
Set the iterator position to the index immediately BEYOND the last character in the text being scanne...
virtual void setText(UText *text, UErrorCode &status)=0
Reset the break iterator to operate over the text represented by the UText.
virtual BreakIterator * clone() const =0
Return a polymorphic copy of this object.
virtual UClassID getDynamicClassID(void) const override=0
Return a polymorphic class ID for this object.
static BreakIterator * createWordInstance(const Locale &where, UErrorCode &status)
Create BreakIterator for word-breaks using the given locale.
const char * getLocaleID(ULocDataLocaleType type, UErrorCode &status) const
Get the locale for this break iterator object.
static UnicodeString & getDisplayName(const Locale &objectLocale, const Locale &displayLocale, UnicodeString &name)
Get name of the object for the desired Locale, in the desired language.
Locale getLocale(ULocDataLocaleType type, UErrorCode &status) const
Returns the locale for this break iterator.
bool operator!=(const BreakIterator &rhs) const
Returns the complement of the result of operator==.
Definition: brkiter.h:135
virtual void setText(const UnicodeString &text)=0
Change the text over which this operates.
static BreakIterator * createTitleInstance(const Locale &where, UErrorCode &status)
Create BreakIterator for title-casing breaks using the specified locale Returns an instance of a Brea...
virtual int32_t next(int32_t n)=0
Set the iterator position to the nth boundary from the current boundary.
virtual int32_t preceding(int32_t offset)=0
Set the iterator position to the first boundary preceding the specified offset.
virtual UBool isBoundary(int32_t offset)=0
Return true if the specified position is a boundary position.
static UnicodeString & getDisplayName(const Locale &objectLocale, UnicodeString &name)
Get name of the object for the desired Locale, in the language of the default locale.
static BreakIterator * createSentenceInstance(const Locale &where, UErrorCode &status)
Create BreakIterator for sentence-breaks using specified locale Returns an instance of a BreakIterato...
virtual BreakIterator & refreshInputText(UText *input, UErrorCode &status)=0
Set the subject text string upon which the break iterator is operating without changing any other asp...
virtual int32_t first(void)=0
Sets the current iteration position to the beginning of the text, position zero.
static UBool unregister(URegistryKey key, UErrorCode &status)
Unregister a previously-registered BreakIterator using the key returned from the register call.
virtual UText * getUText(UText *fillIn, UErrorCode &status) const =0
Get a UText for the text being analyzed.
static StringEnumeration * getAvailableLocales(void)
Return a StringEnumeration over the locales available at the time of the call, including registered l...
static BreakIterator * createLineInstance(const Locale &where, UErrorCode &status)
Create BreakIterator for line-breaks using specified locale.
virtual BreakIterator * createBufferClone(void *stackBuffer, int32_t &BufferSize, UErrorCode &status)=0
Deprecated functionality.
virtual int32_t following(int32_t offset)=0
Advance the iterator to the first boundary following the specified offset.
virtual int32_t previous(void)=0
Set the iterator position to the boundary preceding the current boundary.
static URegistryKey registerInstance(BreakIterator *toAdopt, const Locale &locale, UBreakIteratorType kind, UErrorCode &status)
Register a new break iterator of the indicated kind, to use in the given locale.
static const Locale * getAvailableLocales(int32_t &count)
Get the set of Locales for which TextBoundaries are installed.
virtual int32_t getRuleStatusVec(int32_t *fillInVec, int32_t capacity, UErrorCode &status)
For RuleBasedBreakIterators, get the status (tag) values from the break rule(s) that determined the b...
virtual int32_t current(void) const =0
Return character index of the current iterator position within the text.
Abstract class that defines an API for iteration on text objects.
Definition: chariter.h:361
A Locale object represents a specific geographical, political, or cultural region.
Definition: locid.h:195
Base class for 'pure' C++ implementations of uenum api.
Definition: strenum.h:61
UObject is the common ICU "boilerplate" class.
Definition: uobject.h:223
UnicodeString is a string class that stores Unicode characters directly and provides similar function...
Definition: unistr.h:296
C++ API: Locale ID object.
U_EXPORT UBool operator==(const StringPiece &x, const StringPiece &y)
Global operator == for StringPiece.
C++ API: String Enumeration.
UText struct.
Definition: utext.h:1328
C API: BreakIterator.
UBreakIteratorType
The possible types of text boundaries.
Definition: ubrk.h:102
#define ULOC_FULLNAME_CAPACITY
Useful constant for the maximum size of the whole locale ID (including the terminating NULL and all k...
Definition: uloc.h:264
ULocDataLocaleType
Constants for *_getLocale() Allow user to select whether she wants information on requested,...
Definition: uloc.h:338
int8_t UBool
The ICU boolean type, a signed-byte integer.
Definition: umachine.h:247
C API: Miscellaneous definitions.
const void * URegistryKey
Opaque type returned by registerInstance, registerFactory and unregister for service registration.
Definition: umisc.h:57
C++ API: Unicode String.
C++ API: Common ICU base class UObject.
void * UClassID
UClassID is used to identify classes without using the compiler's RTTI.
Definition: uobject.h:96
C API: Abstract Unicode Text API.
Basic definitions for ICU, for both C and C++ APIs.
UErrorCode
Standard ICU4C error code type, a substitute for exceptions.
Definition: utypes.h:415
#define U_COMMON_API
Set to export library symbols from inside the common library, and to import them from outside.
Definition: utypes.h:300