The Tor Browser / file comparison

comparison: intl/icu/source/common/unicode/rbbi.h

intl/icu/source/common/unicode/rbbi.h

branch
TOR_BUG_3246
changeset 4
fc2d59ddac77
equal deleted inserted replaced
-1:000000000000 0:4c23d5dacef3
1 /*
2 ***************************************************************************
3 * Copyright (C) 1999-2013 International Business Machines Corporation *
4 * and others. All rights reserved. *
5 ***************************************************************************
6
7 **********************************************************************
8 * Date Name Description
9 * 10/22/99 alan Creation.
10 * 11/11/99 rgillam Complete port from Java.
11 **********************************************************************
12 */
13
14 #ifndef RBBI_H
15 #define RBBI_H
16
17 #include "unicode/utypes.h"
18
19 /**
20 * \file
21 * \brief C++ API: Rule Based Break Iterator
22 */
23
24 #if !UCONFIG_NO_BREAK_ITERATION
25
26 #include "unicode/brkiter.h"
27 #include "unicode/udata.h"
28 #include "unicode/parseerr.h"
29 #include "unicode/schriter.h"
30 #include "unicode/uchriter.h"
31
32
33 struct UTrie;
34
35 U_NAMESPACE_BEGIN
36
37 /** @internal */
38 struct RBBIDataHeader;
39 class RuleBasedBreakIteratorTables;
40 class BreakIterator;
41 class RBBIDataWrapper;
42 class UStack;
43 class LanguageBreakEngine;
44 class UnhandledEngine;
45 struct RBBIStateTable;
46
47
48
49
50 /**
51 *
52 * A subclass of BreakIterator whose behavior is specified using a list of rules.
53 * <p>Instances of this class are most commonly created by the factory methods of
54 * BreakIterator::createWordInstance(), BreakIterator::createLineInstance(), etc.,
55 * and then used via the abstract API in class BreakIterator</p>
56 *
57 * <p>See the ICU User Guide for information on Break Iterator Rules.</p>
58 *
59 * <p>This class is not intended to be subclassed. (Class DictionaryBasedBreakIterator
60 * is a subclass, but that relationship is effectively internal to the ICU
61 * implementation. The subclassing interface to RulesBasedBreakIterator is
62 * not part of the ICU API, and may not remain stable.</p>
63 *
64 */
65 class U_COMMON_API RuleBasedBreakIterator : public BreakIterator {
66
67 protected:
68 /**
69 * The UText through which this BreakIterator accesses the text
70 * @internal
71 */
72 UText *fText;
73
74 /**
75 * A character iterator that refers to the same text as the UText, above.
76 * Only included for compatibility with old API, which was based on CharacterIterators.
77 * Value may be adopted from outside, or one of fSCharIter or fDCharIter, below.
78 */
79 CharacterIterator *fCharIter;
80
81 /**
82 * When the input text is provided by a UnicodeString, this will point to
83 * a characterIterator that wraps that data. Needed only for the
84 * implementation of getText(), a backwards compatibility issue.
85 */
86 StringCharacterIterator *fSCharIter;
87
88 /**
89 * When the input text is provided by a UText, this
90 * dummy CharacterIterator over an empty string will
91 * be returned from getText()
92 */
93 UCharCharacterIterator *fDCharIter;
94
95 /**
96 * The rule data for this BreakIterator instance
97 * @internal
98 */
99 RBBIDataWrapper *fData;
100
101 /** Index of the Rule {tag} values for the most recent match.
102 * @internal
103 */
104 int32_t fLastRuleStatusIndex;
105
106 /**
107 * Rule tag value valid flag.
108 * Some iterator operations don't intrinsically set the correct tag value.
109 * This flag lets us lazily compute the value if we are ever asked for it.
110 * @internal
111 */
112 UBool fLastStatusIndexValid;
113
114 /**
115 * Counter for the number of characters encountered with the "dictionary"
116 * flag set.
117 * @internal
118 */
119 uint32_t fDictionaryCharCount;
120
121 /**
122 * When a range of characters is divided up using the dictionary, the break
123 * positions that are discovered are stored here, preventing us from having
124 * to use either the dictionary or the state table again until the iterator
125 * leaves this range of text. Has the most impact for line breaking.
126 * @internal
127 */
128 int32_t* fCachedBreakPositions;
129
130 /**
131 * The number of elements in fCachedBreakPositions
132 * @internal
133 */
134 int32_t fNumCachedBreakPositions;
135
136 /**
137 * if fCachedBreakPositions is not null, this indicates which item in the
138 * cache the current iteration position refers to
139 * @internal
140 */
141 int32_t fPositionInCache;
142
143 /**
144 *
145 * If present, UStack of LanguageBreakEngine objects that might handle
146 * dictionary characters. Searched from top to bottom to find an object to
147 * handle a given character.
148 * @internal
149 */
150 UStack *fLanguageBreakEngines;
151
152 /**
153 *
154 * If present, the special LanguageBreakEngine used for handling
155 * characters that are in the dictionary set, but not handled by any
156 * LangugageBreakEngine.
157 * @internal
158 */
159 UnhandledEngine *fUnhandledBreakEngine;
160
161 /**
162 *
163 * The type of the break iterator, or -1 if it has not been set.
164 * @internal
165 */
166 int32_t fBreakType;
167
168 protected:
169 //=======================================================================
170 // constructors
171 //=======================================================================
172
173 #ifndef U_HIDE_INTERNAL_API
174 /**
175 * Constant to be used in the constructor
176 * RuleBasedBreakIterator(RBBIDataHeader*, EDontAdopt, UErrorCode &);
177 * which does not adopt the memory indicated by the RBBIDataHeader*
178 * parameter.
179 *
180 * @internal
181 */
182 enum EDontAdopt {
183 kDontAdopt
184 };
185
186 /**
187 * Constructor from a flattened set of RBBI data in malloced memory.
188 * RulesBasedBreakIterators built from a custom set of rules
189 * are created via this constructor; the rules are compiled
190 * into memory, then the break iterator is constructed here.
191 *
192 * The break iterator adopts the memory, and will
193 * free it when done.
194 * @internal
195 */
196 RuleBasedBreakIterator(RBBIDataHeader* data, UErrorCode &status);
197
198 /**
199 * Constructor from a flattened set of RBBI data in memory which need not
200 * be malloced (e.g. it may be a memory-mapped file, etc.).
201 *
202 * This version does not adopt the memory, and does not
203 * free it when done.
204 * @internal
205 */
206 RuleBasedBreakIterator(const RBBIDataHeader* data, enum EDontAdopt dontAdopt, UErrorCode &status);
207 #endif /* U_HIDE_INTERNAL_API */
208
209
210 friend class RBBIRuleBuilder;
211 /** @internal */
212 friend class BreakIterator;
213
214
215
216 public:
217
218 /** Default constructor. Creates an empty shell of an iterator, with no
219 * rules or text to iterate over. Object can subsequently be assigned to.
220 * @stable ICU 2.2
221 */
222 RuleBasedBreakIterator();
223
224 /**
225 * Copy constructor. Will produce a break iterator with the same behavior,
226 * and which iterates over the same text, as the one passed in.
227 * @param that The RuleBasedBreakIterator passed to be copied
228 * @stable ICU 2.0
229 */
230 RuleBasedBreakIterator(const RuleBasedBreakIterator& that);
231
232 /**
233 * Construct a RuleBasedBreakIterator from a set of rules supplied as a string.
234 * @param rules The break rules to be used.
235 * @param parseError In the event of a syntax error in the rules, provides the location
236 * within the rules of the problem.
237 * @param status Information on any errors encountered.
238 * @stable ICU 2.2
239 */
240 RuleBasedBreakIterator( const UnicodeString &rules,
241 UParseError &parseError,
242 UErrorCode &status);
243
244 /**
245 * Contruct a RuleBasedBreakIterator from a set of precompiled binary rules.
246 * Binary rules are obtained from RulesBasedBreakIterator::getBinaryRules().
247 * Construction of a break iterator in this way is substantially faster than
248 * constuction from source rules.
249 *
250 * Ownership of the storage containing the compiled rules remains with the
251 * caller of this function. The compiled rules must not be modified or
252 * deleted during the life of the break iterator.
253 *
254 * The compiled rules are not compatible across different major versions of ICU.
255 * The compiled rules are comaptible only between machines with the same
256 * byte ordering (little or big endian) and the same base character set family
257 * (ASCII or EBCDIC).
258 *
259 * @see #getBinaryRules
260 * @param compiledRules A pointer to the compiled break rules to be used.
261 * @param ruleLength The length of the compiled break rules, in bytes. This
262 * corresponds to the length value produced by getBinaryRules().
263 * @param status Information on any errors encountered, including invalid
264 * binary rules.
265 * @stable ICU 4.8
266 */
267 RuleBasedBreakIterator(const uint8_t *compiledRules,
268 uint32_t ruleLength,
269 UErrorCode &status);
270
271 /**
272 * This constructor uses the udata interface to create a BreakIterator
273 * whose internal tables live in a memory-mapped file. "image" is an
274 * ICU UDataMemory handle for the pre-compiled break iterator tables.
275 * @param image handle to the memory image for the break iterator data.
276 * Ownership of the UDataMemory handle passes to the Break Iterator,
277 * which will be responsible for closing it when it is no longer needed.
278 * @param status Information on any errors encountered.
279 * @see udata_open
280 * @see #getBinaryRules
281 * @stable ICU 2.8
282 */
283 RuleBasedBreakIterator(UDataMemory* image, UErrorCode &status);
284
285 /**
286 * Destructor
287 * @stable ICU 2.0
288 */
289 virtual ~RuleBasedBreakIterator();
290
291 /**
292 * Assignment operator. Sets this iterator to have the same behavior,
293 * and iterate over the same text, as the one passed in.
294 * @param that The RuleBasedBreakItertor passed in
295 * @return the newly created RuleBasedBreakIterator
296 * @stable ICU 2.0
297 */
298 RuleBasedBreakIterator& operator=(const RuleBasedBreakIterator& that);
299
300 /**
301 * Equality operator. Returns TRUE if both BreakIterators are of the
302 * same class, have the same behavior, and iterate over the same text.
303 * @param that The BreakIterator to be compared for equality
304 * @return TRUE if both BreakIterators are of the
305 * same class, have the same behavior, and iterate over the same text.
306 * @stable ICU 2.0
307 */
308 virtual UBool operator==(const BreakIterator& that) const;
309
310 /**
311 * Not-equal operator. If operator== returns TRUE, this returns FALSE,
312 * and vice versa.
313 * @param that The BreakIterator to be compared for inequality
314 * @return TRUE if both BreakIterators are not same.
315 * @stable ICU 2.0
316 */
317 UBool operator!=(const BreakIterator& that) const;
318
319 /**
320 * Returns a newly-constructed RuleBasedBreakIterator with the same
321 * behavior, and iterating over the same text, as this one.
322 * Differs from the copy constructor in that it is polymorphic, and
323 * will correctly clone (copy) a derived class.
324 * clone() is thread safe. Multiple threads may simultaeneously
325 * clone the same source break iterator.
326 * @return a newly-constructed RuleBasedBreakIterator
327 * @stable ICU 2.0
328 */
329 virtual BreakIterator* clone() const;
330
331 /**
332 * Compute a hash code for this BreakIterator
333 * @return A hash code
334 * @stable ICU 2.0
335 */
336 virtual int32_t hashCode(void) const;
337
338 /**
339 * Returns the description used to create this iterator
340 * @return the description used to create this iterator
341 * @stable ICU 2.0
342 */
343 virtual const UnicodeString& getRules(void) const;
344
345 //=======================================================================
346 // BreakIterator overrides
347 //=======================================================================
348
349 /**
350 * <p>
351 * Return a CharacterIterator over the text being analyzed.
352 * The returned character iterator is owned by the break iterator, and must
353 * not be deleted by the caller. Repeated calls to this function may
354 * return the same CharacterIterator.
355 * </p>
356 * <p>
357 * The returned character iterator must not be used concurrently with
358 * the break iterator. If concurrent operation is needed, clone the
359 * returned character iterator first and operate on the clone.
360 * </p>
361 * <p>
362 * When the break iterator is operating on text supplied via a UText,
363 * this function will fail. Lacking any way to signal failures, it
364 * returns an CharacterIterator containing no text.
365 * The function getUText() provides similar functionality,
366 * is reliable, and is more efficient.
367 * </p>
368 *
369 * TODO: deprecate this function?
370 *
371 * @return An iterator over the text being analyzed.
372 * @stable ICU 2.0
373 */
374 virtual CharacterIterator& getText(void) const;
375
376
377 /**
378 * Get a UText for the text being analyzed.
379 * The returned UText is a shallow clone of the UText used internally
380 * by the break iterator implementation. It can safely be used to
381 * access the text without impacting any break iterator operations,
382 * but the underlying text itself must not be altered.
383 *
384 * @param fillIn A UText to be filled in. If NULL, a new UText will be
385 * allocated to hold the result.
386 * @param status receives any error codes.
387 * @return The current UText for this break iterator. If an input
388 * UText was provided, it will always be returned.
389 * @stable ICU 3.4
390 */
391 virtual UText *getUText(UText *fillIn, UErrorCode &status) const;
392
393 /**
394 * Set the iterator to analyze a new piece of text. This function resets
395 * the current iteration position to the beginning of the text.
396 * @param newText An iterator over the text to analyze. The BreakIterator
397 * takes ownership of the character iterator. The caller MUST NOT delete it!
398 * @stable ICU 2.0
399 */
400 virtual void adoptText(CharacterIterator* newText);
401
402 /**
403 * Set the iterator to analyze a new piece of text. This function resets
404 * the current iteration position to the beginning of the text.
405 * @param newText The text to analyze.
406 * @stable ICU 2.0
407 */
408 virtual void setText(const UnicodeString& newText);
409
410 /**
411 * Reset the break iterator to operate over the text represented by
412 * the UText. The iterator position is reset to the start.
413 *
414 * This function makes a shallow clone of the supplied UText. This means
415 * that the caller is free to immediately close or otherwise reuse the
416 * Utext that was passed as a parameter, but that the underlying text itself
417 * must not be altered while being referenced by the break iterator.
418 *
419 * @param text The UText used to change the text.
420 * @param status Receives any error codes.
421 * @stable ICU 3.4
422 */
423 virtual void setText(UText *text, UErrorCode &status);
424
425 /**
426 * Sets the current iteration position to the beginning of the text.
427 * @return The offset of the beginning of the text.
428 * @stable ICU 2.0
429 */
430 virtual int32_t first(void);
431
432 /**
433 * Sets the current iteration position to the end of the text.
434 * @return The text's past-the-end offset.
435 * @stable ICU 2.0
436 */
437 virtual int32_t last(void);
438
439 /**
440 * Advances the iterator either forward or backward the specified number of steps.
441 * Negative values move backward, and positive values move forward. This is
442 * equivalent to repeatedly calling next() or previous().
443 * @param n The number of steps to move. The sign indicates the direction
444 * (negative is backwards, and positive is forwards).
445 * @return The character offset of the boundary position n boundaries away from
446 * the current one.
447 * @stable ICU 2.0
448 */
449 virtual int32_t next(int32_t n);
450
451 /**
452 * Advances the iterator to the next boundary position.
453 * @return The position of the first boundary after this one.
454 * @stable ICU 2.0
455 */
456 virtual int32_t next(void);
457
458 /**
459 * Moves the iterator backwards, to the last boundary preceding this one.
460 * @return The position of the last boundary position preceding this one.
461 * @stable ICU 2.0
462 */
463 virtual int32_t previous(void);
464
465 /**
466 * Sets the iterator to refer to the first boundary position following
467 * the specified position.
468 * @param offset The position from which to begin searching for a break position.
469 * @return The position of the first break after the current position.
470 * @stable ICU 2.0
471 */
472 virtual int32_t following(int32_t offset);
473
474 /**
475 * Sets the iterator to refer to the last boundary position before the
476 * specified position.
477 * @param offset The position to begin searching for a break from.
478 * @return The position of the last boundary before the starting position.
479 * @stable ICU 2.0
480 */
481 virtual int32_t preceding(int32_t offset);
482
483 /**
484 * Returns true if the specfied position is a boundary position. As a side
485 * effect, leaves the iterator pointing to the first boundary position at
486 * or after "offset".
487 * @param offset the offset to check.
488 * @return True if "offset" is a boundary position.
489 * @stable ICU 2.0
490 */
491 virtual UBool isBoundary(int32_t offset);
492
493 /**
494 * Returns the current iteration position.
495 * @return The current iteration position.
496 * @stable ICU 2.0
497 */
498 virtual int32_t current(void) const;
499
500
501 /**
502 * Return the status tag from the break rule that determined the most recently
503 * returned break position. For break rules that do not specify a
504 * status, a default value of 0 is returned. If more than one break rule
505 * would cause a boundary to be located at some position in the text,
506 * the numerically largest of the applicable status values is returned.
507 * <p>
508 * Of the standard types of ICU break iterators, only word break and
509 * line break provide status values. The values are defined in
510 * the header file ubrk.h. For Word breaks, the status allows distinguishing between words
511 * that contain alphabetic letters, "words" that appear to be numbers,
512 * punctuation and spaces, words containing ideographic characters, and
513 * more. For Line Break, the status distinguishes between hard (mandatory) breaks
514 * and soft (potential) break positions.
515 * <p>
516 * <code>getRuleStatus()</code> can be called after obtaining a boundary
517 * position from <code>next()</code>, <code>previous()</code>, or
518 * any other break iterator functions that returns a boundary position.
519 * <p>
520 * When creating custom break rules, one is free to define whatever
521 * status values may be convenient for the application.
522 * <p>
523 * Note: this function is not thread safe. It should not have been
524 * declared const, and the const remains only for compatibility
525 * reasons. (The function is logically const, but not bit-wise const).
526 * <p>
527 * @return the status from the break rule that determined the most recently
528 * returned break position.
529 *
530 * @see UWordBreak
531 * @stable ICU 2.2
532 */
533 virtual int32_t getRuleStatus() const;
534
535 /**
536 * Get the status (tag) values from the break rule(s) that determined the most
537 * recently returned break position.
538 * <p>
539 * The returned status value(s) are stored into an array provided by the caller.
540 * The values are stored in sorted (ascending) order.
541 * If the capacity of the output array is insufficient to hold the data,
542 * the output will be truncated to the available length, and a
543 * U_BUFFER_OVERFLOW_ERROR will be signaled.
544 *
545 * @param fillInVec an array to be filled in with the status values.
546 * @param capacity the length of the supplied vector. A length of zero causes
547 * the function to return the number of status values, in the
548 * normal way, without attemtping to store any values.
549 * @param status receives error codes.
550 * @return The number of rule status values from rules that determined
551 * the most recent boundary returned by the break iterator.
552 * In the event of a U_BUFFER_OVERFLOW_ERROR, the return value
553 * is the total number of status values that were available,
554 * not the reduced number that were actually returned.
555 * @see getRuleStatus
556 * @stable ICU 3.0
557 */
558 virtual int32_t getRuleStatusVec(int32_t *fillInVec, int32_t capacity, UErrorCode &status);
559
560 /**
561 * Returns a unique class ID POLYMORPHICALLY. Pure virtual override.
562 * This method is to implement a simple version of RTTI, since not all
563 * C++ compilers support genuine RTTI. Polymorphic operator==() and
564 * clone() methods call this method.
565 *
566 * @return The class ID for this object. All objects of a
567 * given class have the same class ID. Objects of
568 * other classes have different class IDs.
569 * @stable ICU 2.0
570 */
571 virtual UClassID getDynamicClassID(void) const;
572
573 /**
574 * Returns the class ID for this class. This is useful only for
575 * comparing to a return value from getDynamicClassID(). For example:
576 *
577 * Base* polymorphic_pointer = createPolymorphicObject();
578 * if (polymorphic_pointer->getDynamicClassID() ==
579 * Derived::getStaticClassID()) ...
580 *
581 * @return The class ID for all objects of this class.
582 * @stable ICU 2.0
583 */
584 static UClassID U_EXPORT2 getStaticClassID(void);
585
586 /**
587 * Deprecated functionality. Use clone() instead.
588 *
589 * Create a clone (copy) of this break iterator in memory provided
590 * by the caller. The idea is to increase performance by avoiding
591 * a storage allocation. Use of this functoin is NOT RECOMMENDED.
592 * Performance gains are minimal, and correct buffer management is
593 * tricky. Use clone() instead.
594 *
595 * @param stackBuffer The pointer to the memory into which the cloned object
596 * should be placed. If NULL, allocate heap memory
597 * for the cloned object.
598 * @param BufferSize The size of the buffer. If zero, return the required
599 * buffer size, but do not clone the object. If the
600 * size was too small (but not zero), allocate heap
601 * storage for the cloned object.
602 *
603 * @param status Error status. U_SAFECLONE_ALLOCATED_WARNING will be
604 * returned if the the provided buffer was too small, and
605 * the clone was therefore put on the heap.
606 *
607 * @return Pointer to the clone object. This may differ from the stackBuffer
608 * address if the byte alignment of the stack buffer was not suitable
609 * or if the stackBuffer was too small to hold the clone.
610 * @deprecated ICU 52. Use clone() instead.
611 */
612 virtual BreakIterator * createBufferClone(void *stackBuffer,
613 int32_t &BufferSize,
614 UErrorCode &status);
615
616
617 /**
618 * Return the binary form of compiled break rules,
619 * which can then be used to create a new break iterator at some
620 * time in the future. Creating a break iterator from pre-compiled rules
621 * is much faster than building one from the source form of the
622 * break rules.
623 *
624 * The binary data can only be used with the same version of ICU
625 * and on the same platform type (processor endian-ness)
626 *
627 * @param length Returns the length of the binary data. (Out paramter.)
628 *
629 * @return A pointer to the binary (compiled) rule data. The storage
630 * belongs to the RulesBasedBreakIterator object, not the
631 * caller, and must not be modified or deleted.
632 * @stable ICU 4.8
633 */
634 virtual const uint8_t *getBinaryRules(uint32_t &length);
635
636 /**
637 * Set the subject text string upon which the break iterator is operating
638 * without changing any other aspect of the matching state.
639 * The new and previous text strings must have the same content.
640 *
641 * This function is intended for use in environments where ICU is operating on
642 * strings that may move around in memory. It provides a mechanism for notifying
643 * ICU that the string has been relocated, and providing a new UText to access the
644 * string in its new position.
645 *
646 * Note that the break iterator implementation never copies the underlying text
647 * of a string being processed, but always operates directly on the original text
648 * provided by the user. Refreshing simply drops the references to the old text
649 * and replaces them with references to the new.
650 *
651 * Caution: this function is normally used only by very specialized,
652 * system-level code. One example use case is with garbage collection that moves
653 * the text in memory.
654 *
655 * @param input The new (moved) text string.
656 * @param status Receives errors detected by this function.
657 * @return *this
658 *
659 * @stable ICU 49
660 */
661 virtual RuleBasedBreakIterator &refreshInputText(UText *input, UErrorCode &status);
662
663
664 protected:
665 //=======================================================================
666 // implementation
667 //=======================================================================
668 /**
669 * Dumps caches and performs other actions associated with a complete change
670 * in text or iteration position.
671 * @internal
672 */
673 virtual void reset(void);
674
675 #if 0
676 /**
677 * Return true if the category lookup for this char
678 * indicates that it is in the set of dictionary lookup chars.
679 * This function is intended for use by dictionary based break iterators.
680 * @return true if the category lookup for this char
681 * indicates that it is in the set of dictionary lookup chars.
682 * @internal
683 */
684 virtual UBool isDictionaryChar(UChar32);
685
686 /**
687 * Get the type of the break iterator.
688 * @internal
689 */
690 virtual int32_t getBreakType() const;
691 #endif
692
693 /**
694 * Set the type of the break iterator.
695 * @internal
696 */
697 virtual void setBreakType(int32_t type);
698
699 #ifndef U_HIDE_INTERNAL_API
700 /**
701 * Common initialization function, used by constructors and bufferClone.
702 * @internal
703 */
704 void init();
705 #endif /* U_HIDE_INTERNAL_API */
706
707 private:
708
709 /**
710 * This method backs the iterator back up to a "safe position" in the text.
711 * This is a position that we know, without any context, must be a break position.
712 * The various calling methods then iterate forward from this safe position to
713 * the appropriate position to return. (For more information, see the description
714 * of buildBackwardsStateTable() in RuleBasedBreakIterator.Builder.)
715 * @param statetable state table used of moving backwards
716 * @internal
717 */
718 int32_t handlePrevious(const RBBIStateTable *statetable);
719
720 /**
721 * This method is the actual implementation of the next() method. All iteration
722 * vectors through here. This method initializes the state machine to state 1
723 * and advances through the text character by character until we reach the end
724 * of the text or the state machine transitions to state 0. We update our return
725 * value every time the state machine passes through a possible end state.
726 * @param statetable state table used of moving forwards
727 * @internal
728 */
729 int32_t handleNext(const RBBIStateTable *statetable);
730
731 protected:
732
733 #ifndef U_HIDE_INTERNAL_API
734 /**
735 * This is the function that actually implements dictionary-based
736 * breaking. Covering at least the range from startPos to endPos,
737 * it checks for dictionary characters, and if it finds them determines
738 * the appropriate object to deal with them. It may cache found breaks in
739 * fCachedBreakPositions as it goes. It may well also look at text outside
740 * the range startPos to endPos.
741 * If going forward, endPos is the normal Unicode break result, and
742 * if goind in reverse, startPos is the normal Unicode break result
743 * @param startPos The start position of a range of text
744 * @param endPos The end position of a range of text
745 * @param reverse The call is for the reverse direction
746 * @internal
747 */
748 int32_t checkDictionary(int32_t startPos, int32_t endPos, UBool reverse);
749 #endif /* U_HIDE_INTERNAL_API */
750
751 private:
752
753 /**
754 * This function returns the appropriate LanguageBreakEngine for a
755 * given character c.
756 * @param c A character in the dictionary set
757 * @internal
758 */
759 const LanguageBreakEngine *getLanguageBreakEngine(UChar32 c);
760
761 /**
762 * @internal
763 */
764 void makeRuleStatusValid();
765
766 };
767
768 //------------------------------------------------------------------------------
769 //
770 // Inline Functions Definitions ...
771 //
772 //------------------------------------------------------------------------------
773
774 inline UBool RuleBasedBreakIterator::operator!=(const BreakIterator& that) const {
775 return !operator==(that);
776 }
777
778 U_NAMESPACE_END
779
780 #endif /* #if !UCONFIG_NO_BREAK_ITERATION */
781
782 #endif

Mercurial Repository: The Tor Browser

Europalab SCM Repository Server

mercurial