X-Git-Url: https://git.saurik.com/apple/icu.git/blobdiff_plain/b75a7d8f3b4adbae880cab104ce2c6a50eee4db2..refs/heads/master:/icuSources/common/unicode/rbbi.h diff --git a/icuSources/common/unicode/rbbi.h b/icuSources/common/unicode/rbbi.h index fdad980a..84f67852 100644 --- a/icuSources/common/unicode/rbbi.h +++ b/icuSources/common/unicode/rbbi.h @@ -1,6 +1,8 @@ +// © 2016 and later: Unicode, Inc. and others. +// License & terms of use: http://www.unicode.org/copyright.html /* *************************************************************************** -* Copyright (C) 1999-2003 International Business Machines Corporation * +* Copyright (C) 1999-2016 International Business Machines Corporation * * and others. All rights reserved. * *************************************************************************** @@ -16,24 +18,33 @@ #include "unicode/utypes.h" +#if U_SHOW_CPLUSPLUS_API + +/** + * \file + * \brief C++ API: Rule Based Break Iterator + */ + #if !UCONFIG_NO_BREAK_ITERATION #include "unicode/brkiter.h" #include "unicode/udata.h" #include "unicode/parseerr.h" - -struct UTrie; +#include "unicode/schriter.h" +// for Apple addition: +#include "unicode/urbtok.h" U_NAMESPACE_BEGIN +/** @internal */ +class LanguageBreakEngine; struct RBBIDataHeader; -class RuleBasedBreakIteratorTables; -class BreakIterator; class RBBIDataWrapper; - - +class UnhandledEngine; +class UStack; /** + * * A subclass of BreakIterator whose behavior is specified using a list of rules. *

Instances of this class are most commonly created by the factory methods of * BreakIterator::createWordInstance(), BreakIterator::createLineInstance(), etc., @@ -41,78 +52,120 @@ class RBBIDataWrapper; * *

See the ICU User Guide for information on Break Iterator Rules.

* - *

This class is not intended to be subclassed. (Class DictionaryBasedBreakIterator - * is a subclass, but that relationship is effectively internal to the ICU - * implementation. The subclassing interface to RulesBasedBreakIterator is - * not part of the ICU API, and may not remain stable.

- * + *

This class is not intended to be subclassed.

*/ -class U_COMMON_API RuleBasedBreakIterator : public BreakIterator { +class U_COMMON_API RuleBasedBreakIterator /*U_FINAL*/ : public BreakIterator { -protected: +private: /** - * The character iterator through which this BreakIterator accesses the text - * @internal + * The UText through which this BreakIterator accesses the text + * @internal (private) */ - CharacterIterator* fText; + UText fText; +#ifndef U_HIDE_INTERNAL_API +public: +#endif /* U_HIDE_INTERNAL_API */ /** - * The rule data for this BreakIterator instance + * The rule data for this BreakIterator instance. + * Not for general use; Public only for testing purposes. * @internal */ RBBIDataWrapper *fData; - /** @internal */ - UTrie *fCharMappings; +private: - /** Rule {tag} value for the most recent match. - * @internal - */ - int32_t fLastBreakTag; + /** + * Character categories for the Latin1 subset of Unicode + * @internal Apple-only + */ + uint16_t *fLatin1Cat; /** - * Rule tag value valid flag. - * Some iterator operations don't intrinsically set the correct tag value. - * This flag lets us lazily compute the value if we are ever asked for it. - * @internal + * Character category overrides + * @internal Apple-only */ - UBool fLastBreakTagValid; + typedef struct { + UChar32 c; + uint16_t category; + } CategoryOverride; + CategoryOverride *fCatOverrides; + int32_t fCatOverrideCount; /** - * Counter for the number of characters encountered with the "dictionary" - * flag set. Normal RBBI iterators don't use it, although the code - * for updating it is live. Dictionary Based break iterators (a subclass - * of us) access this field directly. - * @internal + * The current position of the iterator. Pinned, 0 < fPosition <= text.length. + * Never has the value UBRK_DONE (-1). + */ + int32_t fPosition; + + /** + * TODO: + */ + int32_t fRuleStatusIndex; + + /** + * Cache of previously determined boundary positions. */ - uint32_t fDictionaryCharCount; + class BreakCache; + BreakCache *fBreakCache; /** - * Debugging flag. Trace operation of state machine when true. - * @internal + * Cache of boundary positions within a region of text that has been + * sub-divided by dictionary based breaking. */ - static UBool fTrace; + class DictionaryCache; + DictionaryCache *fDictionaryCache; + /** + * + * If present, UStack of LanguageBreakEngine objects that might handle + * dictionary characters. Searched from top to bottom to find an object to + * handle a given character. + * @internal (private) + */ + UStack *fLanguageBreakEngines; + /** + * + * If present, the special LanguageBreakEngine used for handling + * characters that are in the dictionary set, but not handled by any + * LanguageBreakEngine. + * @internal (private) + */ + UnhandledEngine *fUnhandledBreakEngine; -private: /** - * Class ID + * Counter for the number of characters encountered with the "dictionary" + * flag set. + * @internal (private) */ - static const char fgClassID; + uint32_t fDictionaryCharCount; -protected: - //======================================================================= - // constructors - //======================================================================= + /** + * A character iterator that refers to the same text as the UText, above. + * Only included for compatibility with old API, which was based on CharacterIterators. + * Value may be adopted from outside, or one of fSCharIter or fDCharIter, below. + */ + CharacterIterator *fCharIter; /** - * This constructor uses the udata interface to create a BreakIterator - * whose internal tables live in a memory-mapped file. "image" is a pointer - * to the beginning of that file. - * @internal + * When the input text is provided by a UnicodeString, this will point to + * a characterIterator that wraps that data. Needed only for the + * implementation of getText(), a backwards compatibility issue. */ - RuleBasedBreakIterator(UDataMemory* image, UErrorCode &status); + StringCharacterIterator fSCharIter; + + /** + * True when iteration has run off the end, and iterator functions should return UBRK_DONE. + */ + UBool fDone; + + //======================================================================= + // constructors + //======================================================================= +// The following is intended to be private in open-source. +// However Apple needs it to be public for urbtok.cpp +public: /** * Constructor from a flattened set of RBBI data in malloced memory. * RulesBasedBreakIterators built from a custom set of rules @@ -121,20 +174,21 @@ protected: * * The break iterator adopts the memory, and will * free it when done. - * @internal + * @internal (private) */ RuleBasedBreakIterator(RBBIDataHeader* data, UErrorCode &status); +private: - friend class RBBIRuleBuilder; /** @internal */ + /** @internal */ + friend class RBBIRuleBuilder; + /** @internal */ friend class BreakIterator; - - public: /** Default constructor. Creates an empty shell of an iterator, with no * rules or text to iterate over. Object can subsequently be assigned to. - * @draft ICU 2.2 + * @stable ICU 2.2 */ RuleBasedBreakIterator(); @@ -152,11 +206,53 @@ public: * @param parseError In the event of a syntax error in the rules, provides the location * within the rules of the problem. * @param status Information on any errors encountered. - * @draft ICU 2.2 + * @stable ICU 2.2 */ RuleBasedBreakIterator( const UnicodeString &rules, UParseError &parseError, UErrorCode &status); + + /** + * Construct a RuleBasedBreakIterator from a set of precompiled binary rules. + * Binary rules are obtained from RulesBasedBreakIterator::getBinaryRules(). + * Construction of a break iterator in this way is substantially faster than + * construction from source rules. + * + * Ownership of the storage containing the compiled rules remains with the + * caller of this function. The compiled rules must not be modified or + * deleted during the life of the break iterator. + * + * The compiled rules are not compatible across different major versions of ICU. + * The compiled rules are compatible only between machines with the same + * byte ordering (little or big endian) and the same base character set family + * (ASCII or EBCDIC). + * + * @see #getBinaryRules + * @param compiledRules A pointer to the compiled break rules to be used. + * @param ruleLength The length of the compiled break rules, in bytes. This + * corresponds to the length value produced by getBinaryRules(). + * @param status Information on any errors encountered, including invalid + * binary rules. + * @stable ICU 4.8 + */ + RuleBasedBreakIterator(const uint8_t *compiledRules, + uint32_t ruleLength, + UErrorCode &status); + + /** + * This constructor uses the udata interface to create a BreakIterator + * whose internal tables live in a memory-mapped file. "image" is an + * ICU UDataMemory handle for the pre-compiled break iterator tables. + * @param image handle to the memory image for the break iterator data. + * Ownership of the UDataMemory handle passes to the Break Iterator, + * which will be responsible for closing it when it is no longer needed. + * @param status Information on any errors encountered. + * @see udata_open + * @see #getBinaryRules + * @stable ICU 2.8 + */ + RuleBasedBreakIterator(UDataMemory* image, UErrorCode &status); + /** * Destructor * @stable ICU 2.0 @@ -176,7 +272,7 @@ public: * Equality operator. Returns TRUE if both BreakIterators are of the * same class, have the same behavior, and iterate over the same text. * @param that The BreakIterator to be compared for equality - * @Return TRUE if both BreakIterators are of the + * @return TRUE if both BreakIterators are of the * same class, have the same behavior, and iterate over the same text. * @stable ICU 2.0 */ @@ -189,18 +285,19 @@ public: * @return TRUE if both BreakIterators are not same. * @stable ICU 2.0 */ - UBool operator!=(const BreakIterator& that) const; + inline UBool operator!=(const BreakIterator& that) const; /** * Returns a newly-constructed RuleBasedBreakIterator with the same * behavior, and iterating over the same text, as this one. * Differs from the copy constructor in that it is polymorphic, and - * will correctly clone (copy) a derived class. - * clone() is thread safe. Multiple threads may simultaeneously + * will correctly clone (copy) a derived class. + * clone() is thread safe. Multiple threads may simultaneously * clone the same source break iterator. - * @stable ICU 2.0 + * @return a newly-constructed RuleBasedBreakIterator + * @stable ICU 2.0 */ - virtual BreakIterator* clone() const; + virtual RuleBasedBreakIterator* clone() const; /** * Compute a hash code for this BreakIterator @@ -221,16 +318,49 @@ public: //======================================================================= /** - * Return a CharacterIterator over the text being analyzed. This version - * of this method returns the actual CharacterIterator we're using internally. - * Changing the state of this iterator can have undefined consequences. If - * you need to change it, clone it first. + *

+ * Return a CharacterIterator over the text being analyzed. + * The returned character iterator is owned by the break iterator, and must + * not be deleted by the caller. Repeated calls to this function may + * return the same CharacterIterator. + *

+ *

+ * The returned character iterator must not be used concurrently with + * the break iterator. If concurrent operation is needed, clone the + * returned character iterator first and operate on the clone. + *

+ *

+ * When the break iterator is operating on text supplied via a UText, + * this function will fail. Lacking any way to signal failures, it + * returns an CharacterIterator containing no text. + * The function getUText() provides similar functionality, + * is reliable, and is more efficient. + *

+ * + * TODO: deprecate this function? + * * @return An iterator over the text being analyzed. - * @stable ICU 2.0 + * @stable ICU 2.0 */ - virtual const CharacterIterator& getText(void) const; + virtual CharacterIterator& getText(void) const; + /** + * Get a UText for the text being analyzed. + * The returned UText is a shallow clone of the UText used internally + * by the break iterator implementation. It can safely be used to + * access the text without impacting any break iterator operations, + * but the underlying text itself must not be altered. + * + * @param fillIn A UText to be filled in. If NULL, a new UText will be + * allocated to hold the result. + * @param status receives any error codes. + * @return The current UText for this break iterator. If an input + * UText was provided, it will always be returned. + * @stable ICU 3.4 + */ + virtual UText *getUText(UText *fillIn, UErrorCode &status) const; + /** * Set the iterator to analyze a new piece of text. This function resets * the current iteration position to the beginning of the text. @@ -243,22 +373,40 @@ public: /** * Set the iterator to analyze a new piece of text. This function resets * the current iteration position to the beginning of the text. + * + * The BreakIterator will retain a reference to the supplied string. + * The caller must not modify or delete the text while the BreakIterator + * retains the reference. + * * @param newText The text to analyze. * @stable ICU 2.0 */ virtual void setText(const UnicodeString& newText); /** - * Sets the current iteration position to the beginning of the text. - * (i.e., the CharacterIterator's starting offset). - * @return The offset of the beginning of the text. + * Reset the break iterator to operate over the text represented by + * the UText. The iterator position is reset to the start. + * + * This function makes a shallow clone of the supplied UText. This means + * that the caller is free to immediately close or otherwise reuse the + * Utext that was passed as a parameter, but that the underlying text itself + * must not be altered while being referenced by the break iterator. + * + * @param text The UText used to change the text. + * @param status Receives any error codes. + * @stable ICU 3.4 + */ + virtual void setText(UText *text, UErrorCode &status); + + /** + * Sets the current iteration position to the beginning of the text, position zero. + * @return The offset of the beginning of the text, zero. * @stable ICU 2.0 */ virtual int32_t first(void); /** * Sets the current iteration position to the end of the text. - * (i.e., the CharacterIterator's ending offset). * @return The text's past-the-end offset. * @stable ICU 2.0 */ @@ -309,7 +457,7 @@ public: virtual int32_t preceding(int32_t offset); /** - * Returns true if the specfied position is a boundary position. As a side + * Returns true if the specified position is a boundary position. As a side * effect, leaves the iterator pointing to the first boundary position at * or after "offset". * @param offset the offset to check. @@ -319,7 +467,10 @@ public: virtual UBool isBoundary(int32_t offset); /** - * Returns the current iteration position. + * Returns the current iteration position. Note that UBRK_DONE is never + * returned from this function; if iteration has run to the end of a + * string, current() will return the length of the string while + * next() will return UBRK_DONE). * @return The current iteration position. * @stable ICU 2.0 */ @@ -327,28 +478,80 @@ public: /** - * Return the status tag from the break rule that determined the most recently - * returned break position. The values appear in the rule source - * within brackets, {123}, for example. For rules that do not specify a - * status, a default value of 0 is returned. + * Return the status tag from the break rule that determined the boundary at + * the current iteration position. For break rules that do not specify a + * status, a default value of 0 is returned. If more than one break rule + * would cause a boundary to be located at some position in the text, + * the numerically largest of the applicable status values is returned. *

- * Of the standard types of ICU break iterators, only the word break - * iterator provides status values. The values are defined in - * enum UWordBreak, and allow distinguishing between words + * Of the standard types of ICU break iterators, only word break and + * line break provide status values. The values are defined in + * the header file ubrk.h. For Word breaks, the status allows distinguishing between words * that contain alphabetic letters, "words" that appear to be numbers, * punctuation and spaces, words containing ideographic characters, and - * more. Call getRuleStatus after obtaining a boundary - * position from next(), previous(), or + * more. For Line Break, the status distinguishes between hard (mandatory) breaks + * and soft (potential) break positions. + *

+ * getRuleStatus() can be called after obtaining a boundary + * position from next(), previous(), or * any other break iterator functions that returns a boundary position. *

- * @return the status from the break rule that determined the most recently - * returned break position. + * Note that getRuleStatus() returns the value corresponding to + * current() index even after next() has returned DONE. + *

+ * When creating custom break rules, one is free to define whatever + * status values may be convenient for the application. + *

+ * @return the status from the break rule that determined the boundary + * at the current iteration position. * * @see UWordBreak - * @draft ICU 2.2 + * @stable ICU 2.2 */ virtual int32_t getRuleStatus() const; + /** + * Get the status (tag) values from the break rule(s) that determined the boundary + * at the current iteration position. + *

+ * The returned status value(s) are stored into an array provided by the caller. + * The values are stored in sorted (ascending) order. + * If the capacity of the output array is insufficient to hold the data, + * the output will be truncated to the available length, and a + * U_BUFFER_OVERFLOW_ERROR will be signaled. + * + * @param fillInVec an array to be filled in with the status values. + * @param capacity the length of the supplied vector. A length of zero causes + * the function to return the number of status values, in the + * normal way, without attempting to store any values. + * @param status receives error codes. + * @return The number of rule status values from the rules that determined + * the boundary at the current iteration position. + * In the event of a U_BUFFER_OVERFLOW_ERROR, the return value + * is the total number of status values that were available, + * not the reduced number that were actually returned. + * @see getRuleStatus + * @stable ICU 3.0 + */ + virtual int32_t getRuleStatusVec(int32_t *fillInVec, int32_t capacity, UErrorCode &status); + + /** + * Apple custom extension + * Initializes Latin1 category + * @internal + */ + void initLatin1Cat(void); + + /** + * Apple custom extension + * Fetch the next set of tokens. + * @param maxTokens The maximum number of tokens to return. + * @param outTokenRanges Pointer to output array of token ranges. + * @param outTokenFlags (optional) pointer to output array of token flags. + * @internal + */ + int32_t tokenize(int32_t maxTokens, RuleBasedTokenRange *outTokenRanges, unsigned long *outTokenFlags); + /** * Returns a unique class ID POLYMORPHICALLY. Pure virtual override. * This method is to implement a simple version of RTTI, since not all @@ -360,7 +563,7 @@ public: * other classes have different class IDs. * @stable ICU 2.0 */ - inline virtual UClassID getDynamicClassID(void) const; + virtual UClassID getDynamicClassID(void) const; /** * Returns the class ID for this class. This is useful only for @@ -373,12 +576,15 @@ public: * @return The class ID for all objects of this class. * @stable ICU 2.0 */ - inline static UClassID getStaticClassID(void); + static UClassID U_EXPORT2 getStaticClassID(void); - /* +#ifndef U_FORCE_HIDE_DEPRECATED_API + /** + * Deprecated functionality. Use clone() instead. + * * Create a clone (copy) of this break iterator in memory provided * by the caller. The idea is to increase performance by avoiding - * a storage allocation. Use of this functoin is NOT RECOMMENDED. + * a storage allocation. Use of this function is NOT RECOMMENDED. * Performance gains are minimal, and correct buffer management is * tricky. Use clone() instead. * @@ -391,18 +597,18 @@ public: * storage for the cloned object. * * @param status Error status. U_SAFECLONE_ALLOCATED_WARNING will be - * returned if the the provided buffer was too small, and + * returned if the provided buffer was too small, and * the clone was therefore put on the heap. * * @return Pointer to the clone object. This may differ from the stackBuffer * address if the byte alignment of the stack buffer was not suitable * or if the stackBuffer was too small to hold the clone. - * @stable ICU 2.0 + * @deprecated ICU 52. Use clone() instead. */ - virtual BreakIterator * createBufferClone(void *stackBuffer, - int32_t &BufferSize, - UErrorCode &status); - + virtual RuleBasedBreakIterator *createBufferClone(void *stackBuffer, + int32_t &BufferSize, + UErrorCode &status); +#endif // U_FORCE_HIDE_DEPRECATED_API /** * Return the binary form of compiled break rules, @@ -411,90 +617,137 @@ public: * is much faster than building one from the source form of the * break rules. * - * The binary data is can only be used with the same version of ICU + * The binary data can only be used with the same version of ICU * and on the same platform type (processor endian-ness) * - * @param length Returns the length of the binary data. (Out paramter.) + * @param length Returns the length of the binary data. (Out parameter.) * * @return A pointer to the binary (compiled) rule data. The storage * belongs to the RulesBasedBreakIterator object, not the * caller, and must not be modified or deleted. - * @internal + * @stable ICU 4.8 */ virtual const uint8_t *getBinaryRules(uint32_t &length); + /** + * Set the subject text string upon which the break iterator is operating + * without changing any other aspect of the matching state. + * The new and previous text strings must have the same content. + * + * This function is intended for use in environments where ICU is operating on + * strings that may move around in memory. It provides a mechanism for notifying + * ICU that the string has been relocated, and providing a new UText to access the + * string in its new position. + * + * Note that the break iterator implementation never copies the underlying text + * of a string being processed, but always operates directly on the original text + * provided by the user. Refreshing simply drops the references to the old text + * and replaces them with references to the new. + * + * Caution: this function is normally used only by very specialized, + * system-level code. One example use case is with garbage collection that moves + * the text in memory. + * + * @param input The new (moved) text string. + * @param status Receives errors detected by this function. + * @return *this + * + * @stable ICU 49 + */ + virtual RuleBasedBreakIterator &refreshInputText(UText *input, UErrorCode &status); + +#ifndef U_HIDE_INTERNAL_API + /** + * Set the break category overrides for this break iterator, based on delimiter data. + * @param locale The locale whose delimiters to use. + * @internal Apple only + */ + virtual void setCategoryOverrides(Locale locale); +#endif /* U_HIDE_INTERNAL_API */ -protected: +private: //======================================================================= // implementation //======================================================================= /** - * This method is the actual implementation of the next() method. All iteration - * vectors through here. This method initializes the state machine to state 1 - * and advances through the text character by character until we reach the end - * of the text or the state machine transitions to state 0. We update our return - * value every time the state machine passes through a possible end state. - * @internal + * Dumps caches and performs other actions associated with a complete change + * in text or iteration position. + * @internal (private) */ - virtual int32_t handleNext(void); + void reset(void); /** - * This method backs the iterator back up to a "safe position" in the text. - * This is a position that we know, without any context, must be a break position. - * The various calling methods then iterate forward from this safe position to - * the appropriate position to return. (For more information, see the description - * of buildBackwardsStateTable() in RuleBasedBreakIterator.Builder.) - * @internal + * Common initialization function, used by constructors and bufferClone. + * @internal (private) + */ + void init(UErrorCode &status); + + /** + * Iterate backwards from an arbitrary position in the input text using the + * synthesized Safe Reverse rules. + * This locates a "Safe Position" from which the forward break rules + * will operate correctly. A Safe Position is not necessarily a boundary itself. + * + * @param fromPosition the position in the input text to begin the iteration. + * @internal (private) */ - virtual int32_t handlePrevious(void); + int32_t handleSafePrevious(int32_t fromPosition); /** - * Dumps caches and performs other actions associated with a complete change - * in text or iteration position. This function is a no-op in RuleBasedBreakIterator, - * but subclasses can and do override it. - * @internal + * Find a rule-based boundary by running the state machine. + * Input + * fPosition, the position in the text to begin from. + * Output + * fPosition: the boundary following the starting position. + * fDictionaryCharCount the number of dictionary characters encountered. + * If > 0, the segment will be further subdivided + * fRuleStatusIndex Info from the state table indicating which rules caused the boundary. + * + * @internal (private) */ - virtual void reset(void); + int32_t handleNext(); + int32_t handleNextInternal(); + /** - * Return true if the category lookup for this char - * indicates that it is in the set of dictionary lookup chars. - * This function is intended for use by dictionary based break iterators. - * @return true if the category lookup for this char - * indicates that it is in the set of dictionary lookup chars. - * @internal - */ - virtual UBool isDictionaryChar(UChar32); + * This function returns the appropriate LanguageBreakEngine for a + * given character c. + * @param c A character in the dictionary set + * @internal (private) + */ + const LanguageBreakEngine *getLanguageBreakEngine(UChar32 c); + public: +#ifndef U_HIDE_INTERNAL_API /** - * Common initialization function, used by constructors and bufferClone. - * (Also used by DictionaryBasedBreakIterator::createBufferClone().) - * @internal - */ - void init(); + * Debugging function only. + * @internal + */ + void dumpCache(); + /** + * Debugging function only. + * @internal + */ + void dumpTables(); + +#endif /* U_HIDE_INTERNAL_API */ }; -//---------------------------------------------------------------------------------- +//------------------------------------------------------------------------------ // // Inline Functions Definitions ... // -//---------------------------------------------------------------------------------- +//------------------------------------------------------------------------------ inline UBool RuleBasedBreakIterator::operator!=(const BreakIterator& that) const { return !operator==(that); } -inline UClassID RuleBasedBreakIterator::getStaticClassID(void) { - return (UClassID)(&fgClassID); -} - -inline UClassID RuleBasedBreakIterator::getDynamicClassID(void) const { - return RuleBasedBreakIterator::getStaticClassID(); -} - U_NAMESPACE_END #endif /* #if !UCONFIG_NO_BREAK_ITERATION */ +#endif /* U_SHOW_CPLUSPLUS_API */ + #endif