X-Git-Url: https://git.saurik.com/apple/icu.git/blobdiff_plain/f59164e3d128c7675a4d3934206346a3384e53a5..f3c0d7a59d99c2a94c6b8822291f0e42be3773c9:/icuSources/common/unicode/ucharstrie.h diff --git a/icuSources/common/unicode/ucharstrie.h b/icuSources/common/unicode/ucharstrie.h index 0575a974..533e8c55 100644 --- a/icuSources/common/unicode/ucharstrie.h +++ b/icuSources/common/unicode/ucharstrie.h @@ -1,10 +1,12 @@ +// © 2016 and later: Unicode, Inc. and others. +// License & terms of use: http://www.unicode.org/copyright.html /* ******************************************************************************* * Copyright (C) 2010-2012, International Business Machines * Corporation and others. All Rights Reserved. ******************************************************************************* * file name: ucharstrie.h -* encoding: US-ASCII +* encoding: UTF-8 * tab size: 8 (not used) * indentation:4 * @@ -26,6 +28,7 @@ #include "unicode/uobject.h" #include "unicode/ustringtrie.h" +#if U_SHOW_CPLUSPLUS_API U_NAMESPACE_BEGIN class Appendable; @@ -34,7 +37,7 @@ class UVector32; /** * Light-weight, non-const reader class for a UCharsTrie. - * Traverses a UChar-serialized data structure with minimal state, + * Traverses a char16_t-serialized data structure with minimal state, * for mapping strings (16-bit-unit sequences) to non-negative integer values. * * This class owns the serialized trie data only if it was constructed by @@ -50,18 +53,18 @@ public: /** * Constructs a UCharsTrie reader instance. * - * The trieUChars must contain a copy of a UChar sequence from the UCharsTrieBuilder, - * starting with the first UChar of that sequence. - * The UCharsTrie object will not read more UChars than + * The trieUChars must contain a copy of a char16_t sequence from the UCharsTrieBuilder, + * starting with the first char16_t of that sequence. + * The UCharsTrie object will not read more char16_ts than * the UCharsTrieBuilder generated in the corresponding build() call. * * The array is not copied/cloned and must not be modified while * the UCharsTrie object is in use. * - * @param trieUChars The UChar array that contains the serialized trie. + * @param trieUChars The char16_t array that contains the serialized trie. * @stable ICU 4.8 */ - UCharsTrie(const UChar *trieUChars) + UCharsTrie(ConstChar16Ptr trieUChars) : ownedArray_(NULL), uchars_(trieUChars), pos_(uchars_), remainingMatchLength_(-1) {} @@ -73,7 +76,7 @@ public: /** * Copy constructor, copies the other trie reader object and its state, - * but not the UChar array which will be shared. (Shallow copy.) + * but not the char16_t array which will be shared. (Shallow copy.) * @param other Another UCharsTrie object. * @stable ICU 4.8 */ @@ -107,8 +110,8 @@ public: private: friend class UCharsTrie; - const UChar *uchars; - const UChar *pos; + const char16_t *uchars; + const char16_t *pos; int32_t remainingMatchLength; }; @@ -146,14 +149,14 @@ public: /** * Determines whether the string so far matches, whether it has a value, - * and whether another input UChar can continue a matching string. + * and whether another input char16_t can continue a matching string. * @return The match/value Result. * @stable ICU 4.8 */ UStringTrieResult current() const; /** - * Traverses the trie from the initial state for this input UChar. + * Traverses the trie from the initial state for this input char16_t. * Equivalent to reset().next(uchar). * @param uchar Input char value. Values below 0 and above 0xffff will never match. * @return The match/value Result. @@ -175,7 +178,7 @@ public: UStringTrieResult firstForCodePoint(UChar32 cp); /** - * Traverses the trie from the current state for this input UChar. + * Traverses the trie from the current state for this input char16_t. * @param uchar Input char value. Values below 0 and above 0xffff will never match. * @return The match/value Result. * @stable ICU 4.8 @@ -206,7 +209,7 @@ public: * @return The match/value Result. * @stable ICU 4.8 */ - UStringTrieResult next(const UChar *s, int32_t length); + UStringTrieResult next(ConstChar16Ptr s, int32_t length); /** * Returns a matching string's value if called immediately after @@ -218,7 +221,7 @@ public: * @stable ICU 4.8 */ inline int32_t getValue() const { - const UChar *pos=pos_; + const char16_t *pos=pos_; int32_t leadUnit=*pos++; // U_ASSERT(leadUnit>=kMinValueLead); return leadUnit&kValueIsFinal ? @@ -235,16 +238,16 @@ public: * @stable ICU 4.8 */ inline UBool hasUniqueValue(int32_t &uniqueValue) const { - const UChar *pos=pos_; + const char16_t *pos=pos_; // Skip the rest of a pending linear-match node. return pos!=NULL && findUniqueValue(pos+remainingMatchLength_+1, FALSE, uniqueValue); } /** - * Finds each UChar which continues the string from the current state. - * That is, each UChar c for which it would be next(c)!=USTRINGTRIE_NO_MATCH now. - * @param out Each next UChar is appended to this object. - * @return the number of UChars which continue the string from here + * Finds each char16_t which continues the string from the current state. + * That is, each char16_t c for which it would be next(c)!=USTRINGTRIE_NO_MATCH now. + * @param out Each next char16_t is appended to this object. + * @return the number of char16_ts which continue the string from here * @stable ICU 4.8 */ int32_t getNextUChars(Appendable &out) const; @@ -256,8 +259,8 @@ public: class U_COMMON_API Iterator : public UMemory { public: /** - * Iterates from the root of a UChar-serialized UCharsTrie. - * @param trieUChars The trie UChars. + * Iterates from the root of a char16_t-serialized UCharsTrie. + * @param trieUChars The trie char16_ts. * @param maxStringLength If 0, the iterator returns full strings. * Otherwise, the iterator returns strings with this maximum length. * @param errorCode Standard ICU error code. Its input value must @@ -266,7 +269,7 @@ public: * function chaining. (See User Guide for details.) * @stable ICU 4.8 */ - Iterator(const UChar *trieUChars, int32_t maxStringLength, UErrorCode &errorCode); + Iterator(ConstChar16Ptr trieUChars, int32_t maxStringLength, UErrorCode &errorCode); /** * Iterates from the current state of the specified UCharsTrie. @@ -334,11 +337,11 @@ public: return TRUE; } - const UChar *branchNext(const UChar *pos, int32_t length, UErrorCode &errorCode); + const char16_t *branchNext(const char16_t *pos, int32_t length, UErrorCode &errorCode); - const UChar *uchars_; - const UChar *pos_; - const UChar *initialPos_; + const char16_t *uchars_; + const char16_t *pos_; + const char16_t *initialPos_; int32_t remainingMatchLength_; int32_t initialRemainingMatchLength_; UBool skipValue_; // Skip intermediate value which was already delivered. @@ -366,7 +369,7 @@ private: * this constructor adopts the builder's array. * This constructor is only called by the builder. */ - UCharsTrie(UChar *adoptUChars, const UChar *trieUChars) + UCharsTrie(char16_t *adoptUChars, const char16_t *trieUChars) : ownedArray_(adoptUChars), uchars_(trieUChars), pos_(uchars_), remainingMatchLength_(-1) {} @@ -379,7 +382,7 @@ private: // Reads a compact 32-bit integer. // pos is already after the leadUnit, and the lead unit has bit 15 reset. - static inline int32_t readValue(const UChar *pos, int32_t leadUnit) { + static inline int32_t readValue(const char16_t *pos, int32_t leadUnit) { int32_t value; if(leadUnit=kMinTwoUnitValueLead) { if(leadUnit=kMinTwoUnitNodeValueLead) { if(leadUnit=kMinTwoUnitDeltaLead) { if(delta==kThreeUnitDeltaLead) { @@ -442,7 +445,7 @@ private: return pos+delta; } - static const UChar *skipDelta(const UChar *pos) { + static const char16_t *skipDelta(const char16_t *pos) { int32_t delta=*pos++; if(delta>=kMinTwoUnitDeltaLead) { if(delta==kThreeUnitDeltaLead) { @@ -459,28 +462,28 @@ private: } // Handles a branch node for both next(uchar) and next(string). - UStringTrieResult branchNext(const UChar *pos, int32_t length, int32_t uchar); + UStringTrieResult branchNext(const char16_t *pos, int32_t length, int32_t uchar); // Requires remainingLength_<0. - UStringTrieResult nextImpl(const UChar *pos, int32_t uchar); + UStringTrieResult nextImpl(const char16_t *pos, int32_t uchar); // Helper functions for hasUniqueValue(). // Recursively finds a unique value (or whether there is not a unique one) // from a branch. - static const UChar *findUniqueValueFromBranch(const UChar *pos, int32_t length, + static const char16_t *findUniqueValueFromBranch(const char16_t *pos, int32_t length, UBool haveUniqueValue, int32_t &uniqueValue); // Recursively finds a unique value (or whether there is not a unique one) // starting from a position on a node lead unit. - static UBool findUniqueValue(const UChar *pos, UBool haveUniqueValue, int32_t &uniqueValue); + static UBool findUniqueValue(const char16_t *pos, UBool haveUniqueValue, int32_t &uniqueValue); // Helper functions for getNextUChars(). // getNextUChars() when pos is on a branch node. - static void getNextBranchUChars(const UChar *pos, int32_t length, Appendable &out); + static void getNextBranchUChars(const char16_t *pos, int32_t length, Appendable &out); // UCharsTrie data structure // - // The trie consists of a series of UChar-serialized nodes for incremental - // Unicode string/UChar sequence matching. (UChar=16-bit unsigned integer) + // The trie consists of a series of char16_t-serialized nodes for incremental + // Unicode string/char16_t sequence matching. (char16_t=16-bit unsigned integer) // The root node is at the beginning of the trie data. // // Types of nodes are distinguished by their node lead unit ranges. @@ -489,9 +492,9 @@ private: // // Node types: // - Final-value node: Stores a 32-bit integer in a compact, variable-length format. - // The value is for the string/UChar sequence so far. + // The value is for the string/char16_t sequence so far. // - Match node, optionally with an intermediate value in a different compact format. - // The value, if present, is for the string/UChar sequence so far. + // The value, if present, is for the string/char16_t sequence so far. // // Aside from the value, which uses the node lead unit's high bits: // @@ -558,19 +561,20 @@ private: static const int32_t kMaxTwoUnitDelta=((kThreeUnitDeltaLead-kMinTwoUnitDeltaLead)<<16)-1; // 0x03feffff - UChar *ownedArray_; + char16_t *ownedArray_; // Fixed value referencing the UCharsTrie words. - const UChar *uchars_; + const char16_t *uchars_; // Iterator variables. // Pointer to next trie unit to read. NULL if no more matches. - const UChar *pos_; + const char16_t *pos_; // Remaining length of a linear-match node, minus 1. Negative if not in such a node. int32_t remainingMatchLength_; }; U_NAMESPACE_END +#endif // U_SHOW_CPLUSPLUS_API #endif // __UCHARSTRIE_H__