Victor Chang | 7322950 | 2020-09-17 13:39:19 +0100 | [diff] [blame] | 1 | // © 2016 and later: Unicode, Inc. and others. |
| 2 | // License & terms of use: http://www.unicode.org/copyright.html |
| 3 | /* |
| 4 | ******************************************************************************* |
| 5 | * Copyright (C) 2013-2014, International Business Machines |
| 6 | * Corporation and others. All Rights Reserved. |
| 7 | ******************************************************************************* |
| 8 | * collationsets.h |
| 9 | * |
| 10 | * created on: 2013feb09 |
| 11 | * created by: Markus W. Scherer |
| 12 | */ |
| 13 | |
| 14 | #ifndef __COLLATIONSETS_H__ |
| 15 | #define __COLLATIONSETS_H__ |
| 16 | |
| 17 | #include "unicode/utypes.h" |
| 18 | |
| 19 | #if !UCONFIG_NO_COLLATION |
| 20 | |
| 21 | #include "unicode/uniset.h" |
| 22 | #include "collation.h" |
| 23 | |
| 24 | U_NAMESPACE_BEGIN |
| 25 | |
| 26 | struct CollationData; |
| 27 | |
| 28 | /** |
| 29 | * Finds the set of characters and strings that sort differently in the tailoring |
| 30 | * from the base data. |
| 31 | * |
| 32 | * Every mapping in the tailoring needs to be compared to the base, |
| 33 | * because some mappings are copied for optimization, and |
| 34 | * all contractions for a character are copied if any contractions for that character |
| 35 | * are added, modified or removed. |
| 36 | * |
| 37 | * It might be simpler to re-parse the rule string, but: |
| 38 | * - That would require duplicating some of the from-rules builder code. |
| 39 | * - That would make the runtime code depend on the builder. |
| 40 | * - That would only work if we have the rule string, and we allow users to |
| 41 | * omit the rule string from data files. |
| 42 | */ |
| 43 | class TailoredSet : public UMemory { |
| 44 | public: |
| 45 | TailoredSet(UnicodeSet *t) |
| 46 | : data(NULL), baseData(NULL), |
| 47 | tailored(t), |
| 48 | suffix(NULL), |
| 49 | errorCode(U_ZERO_ERROR) {} |
| 50 | |
| 51 | void forData(const CollationData *d, UErrorCode &errorCode); |
| 52 | |
| 53 | /** |
| 54 | * @return U_SUCCESS(errorCode) in C++, void in Java |
| 55 | * @internal only public for access by callback |
| 56 | */ |
| 57 | UBool handleCE32(UChar32 start, UChar32 end, uint32_t ce32); |
| 58 | |
| 59 | private: |
| 60 | void compare(UChar32 c, uint32_t ce32, uint32_t baseCE32); |
| 61 | void comparePrefixes(UChar32 c, const UChar *p, const UChar *q); |
| 62 | void compareContractions(UChar32 c, const UChar *p, const UChar *q); |
| 63 | |
| 64 | void addPrefixes(const CollationData *d, UChar32 c, const UChar *p); |
| 65 | void addPrefix(const CollationData *d, const UnicodeString &pfx, UChar32 c, uint32_t ce32); |
| 66 | void addContractions(UChar32 c, const UChar *p); |
| 67 | void addSuffix(UChar32 c, const UnicodeString &sfx); |
| 68 | void add(UChar32 c); |
| 69 | |
| 70 | /** Prefixes are reversed in the data structure. */ |
| 71 | void setPrefix(const UnicodeString &pfx) { |
| 72 | unreversedPrefix = pfx; |
| 73 | unreversedPrefix.reverse(); |
| 74 | } |
| 75 | void resetPrefix() { |
| 76 | unreversedPrefix.remove(); |
| 77 | } |
| 78 | |
| 79 | const CollationData *data; |
| 80 | const CollationData *baseData; |
| 81 | UnicodeSet *tailored; |
| 82 | UnicodeString unreversedPrefix; |
| 83 | const UnicodeString *suffix; |
| 84 | UErrorCode errorCode; |
| 85 | }; |
| 86 | |
| 87 | class ContractionsAndExpansions : public UMemory { |
| 88 | public: |
| 89 | class CESink : public UMemory { |
| 90 | public: |
| 91 | virtual ~CESink(); |
| 92 | virtual void handleCE(int64_t ce) = 0; |
| 93 | virtual void handleExpansion(const int64_t ces[], int32_t length) = 0; |
| 94 | }; |
| 95 | |
| 96 | ContractionsAndExpansions(UnicodeSet *con, UnicodeSet *exp, CESink *s, UBool prefixes) |
| 97 | : data(NULL), |
| 98 | contractions(con), expansions(exp), |
| 99 | sink(s), |
| 100 | addPrefixes(prefixes), |
| 101 | checkTailored(0), |
| 102 | suffix(NULL), |
| 103 | errorCode(U_ZERO_ERROR) {} |
| 104 | |
| 105 | void forData(const CollationData *d, UErrorCode &errorCode); |
| 106 | void forCodePoint(const CollationData *d, UChar32 c, UErrorCode &ec); |
| 107 | |
| 108 | // all following: @internal, only public for access by callback |
| 109 | |
| 110 | void handleCE32(UChar32 start, UChar32 end, uint32_t ce32); |
| 111 | |
| 112 | void handlePrefixes(UChar32 start, UChar32 end, uint32_t ce32); |
| 113 | void handleContractions(UChar32 start, UChar32 end, uint32_t ce32); |
| 114 | |
| 115 | void addExpansions(UChar32 start, UChar32 end); |
| 116 | void addStrings(UChar32 start, UChar32 end, UnicodeSet *set); |
| 117 | |
| 118 | /** Prefixes are reversed in the data structure. */ |
| 119 | void setPrefix(const UnicodeString &pfx) { |
| 120 | unreversedPrefix = pfx; |
| 121 | unreversedPrefix.reverse(); |
| 122 | } |
| 123 | void resetPrefix() { |
| 124 | unreversedPrefix.remove(); |
| 125 | } |
| 126 | |
| 127 | const CollationData *data; |
| 128 | UnicodeSet *contractions; |
| 129 | UnicodeSet *expansions; |
| 130 | CESink *sink; |
| 131 | UBool addPrefixes; |
| 132 | int8_t checkTailored; // -1: collected tailored +1: exclude tailored |
| 133 | UnicodeSet tailored; |
| 134 | UnicodeSet ranges; |
| 135 | UnicodeString unreversedPrefix; |
| 136 | const UnicodeString *suffix; |
| 137 | int64_t ces[Collation::MAX_EXPANSION_LENGTH]; |
| 138 | UErrorCode errorCode; |
| 139 | }; |
| 140 | |
| 141 | U_NAMESPACE_END |
| 142 | |
| 143 | #endif // !UCONFIG_NO_COLLATION |
| 144 | #endif // __COLLATIONSETS_H__ |