intl/icu/source/i18n/rbt_data.h@fc2d59ddac77 (annotated)
intl/icu/source/i18n/rbt_data.h
Wed, 31 Dec 2014 07:22:50 +0100
- author
- Michael Schloh von Bennewitz <michael@schloh.com>
- date
- Wed, 31 Dec 2014 07:22:50 +0100
- branch
- TOR_BUG_3246
- changeset 4
- fc2d59ddac77
- permissions
- -rw-r--r--
Correct previous dual key logic pending first delivery installment.
| michael@0 | 1 | /* |
| michael@0 | 2 | ********************************************************************** |
| michael@0 | 3 | * Copyright (C) 1999-2007, International Business Machines Corporation |
| michael@0 | 4 | * and others. All Rights Reserved. |
| michael@0 | 5 | ********************************************************************** |
| michael@0 | 6 | * Date Name Description |
| michael@0 | 7 | * 11/17/99 aliu Creation. |
| michael@0 | 8 | ********************************************************************** |
| michael@0 | 9 | */ |
| michael@0 | 10 | #ifndef RBT_DATA_H |
| michael@0 | 11 | #define RBT_DATA_H |
| michael@0 | 12 | |
| michael@0 | 13 | #include "unicode/utypes.h" |
| michael@0 | 14 | #include "unicode/uclean.h" |
| michael@0 | 15 | |
| michael@0 | 16 | #if !UCONFIG_NO_TRANSLITERATION |
| michael@0 | 17 | |
| michael@0 | 18 | #include "unicode/uobject.h" |
| michael@0 | 19 | #include "rbt_set.h" |
| michael@0 | 20 | #include "hash.h" |
| michael@0 | 21 | |
| michael@0 | 22 | U_NAMESPACE_BEGIN |
| michael@0 | 23 | |
| michael@0 | 24 | class UnicodeFunctor; |
| michael@0 | 25 | class UnicodeMatcher; |
| michael@0 | 26 | class UnicodeReplacer; |
| michael@0 | 27 | |
| michael@0 | 28 | /** |
| michael@0 | 29 | * The rule data for a RuleBasedTransliterators. RBT objects hold |
| michael@0 | 30 | * a const pointer to a TRD object that they do not own. TRD objects |
| michael@0 | 31 | * are essentially the parsed rules in compact, usable form. The |
| michael@0 | 32 | * TRD objects themselves are held for the life of the process in |
| michael@0 | 33 | * a static cache owned by Transliterator. |
| michael@0 | 34 | * |
| michael@0 | 35 | * This class' API is a little asymmetric. There is a method to |
| michael@0 | 36 | * define a variable, but no way to define a set. This is because the |
| michael@0 | 37 | * sets are defined by the parser in a UVector, and the vector is |
| michael@0 | 38 | * copied into a fixed-size array here. Once this is done, no new |
| michael@0 | 39 | * sets may be defined. In practice, there is no need to do so, since |
| michael@0 | 40 | * generating the data and using it are discrete phases. When there |
| michael@0 | 41 | * is a need to access the set data during the parse phase, another |
| michael@0 | 42 | * data structure handles this. See the parsing code for more |
| michael@0 | 43 | * details. |
| michael@0 | 44 | */ |
| michael@0 | 45 | class TransliterationRuleData : public UMemory { |
| michael@0 | 46 | |
| michael@0 | 47 | public: |
| michael@0 | 48 | |
| michael@0 | 49 | // PUBLIC DATA MEMBERS |
| michael@0 | 50 | |
| michael@0 | 51 | /** |
| michael@0 | 52 | * Rule table. May be empty. |
| michael@0 | 53 | */ |
| michael@0 | 54 | TransliterationRuleSet ruleSet; |
| michael@0 | 55 | |
| michael@0 | 56 | /** |
| michael@0 | 57 | * Map variable name (String) to variable (UnicodeString). A variable name |
| michael@0 | 58 | * corresponds to zero or more characters, stored in a UnicodeString in |
| michael@0 | 59 | * this hash. One or more of these chars may also correspond to a |
| michael@0 | 60 | * UnicodeMatcher, in which case the character in the UnicodeString in this hash is |
| michael@0 | 61 | * a stand-in: it is an index for a secondary lookup in |
| michael@0 | 62 | * data.variables. The stand-in also represents the UnicodeMatcher in |
| michael@0 | 63 | * the stored rules. |
| michael@0 | 64 | */ |
| michael@0 | 65 | Hashtable variableNames; |
| michael@0 | 66 | |
| michael@0 | 67 | /** |
| michael@0 | 68 | * Map category variable (UChar) to set (UnicodeFunctor). |
| michael@0 | 69 | * Variables that correspond to a set of characters are mapped |
| michael@0 | 70 | * from variable name to a stand-in character in data.variableNames. |
| michael@0 | 71 | * The stand-in then serves as a key in this hash to lookup the |
| michael@0 | 72 | * actual UnicodeFunctor object. In addition, the stand-in is |
| michael@0 | 73 | * stored in the rule text to represent the set of characters. |
| michael@0 | 74 | * variables[i] represents character (variablesBase + i). |
| michael@0 | 75 | */ |
| michael@0 | 76 | UnicodeFunctor** variables; |
| michael@0 | 77 | |
| michael@0 | 78 | /** |
| michael@0 | 79 | * Flag that indicates whether the variables are owned (if a single |
| michael@0 | 80 | * call to Transliterator::createFromRules() produces a CompoundTransliterator |
| michael@0 | 81 | * with more than one RuleBasedTransliterator as children, they all share |
| michael@0 | 82 | * the same variables list, so only the first one is considered to own |
| michael@0 | 83 | * the variables) |
| michael@0 | 84 | */ |
| michael@0 | 85 | UBool variablesAreOwned; |
| michael@0 | 86 | |
| michael@0 | 87 | /** |
| michael@0 | 88 | * The character that represents variables[0]. Characters |
| michael@0 | 89 | * variablesBase through variablesBase + |
| michael@0 | 90 | * variablesLength - 1 represent UnicodeFunctor objects. |
| michael@0 | 91 | */ |
| michael@0 | 92 | UChar variablesBase; |
| michael@0 | 93 | |
| michael@0 | 94 | /** |
| michael@0 | 95 | * The length of variables. |
| michael@0 | 96 | */ |
| michael@0 | 97 | int32_t variablesLength; |
| michael@0 | 98 | |
| michael@0 | 99 | public: |
| michael@0 | 100 | |
| michael@0 | 101 | /** |
| michael@0 | 102 | * Constructor |
| michael@0 | 103 | * @param status Output param set to success/failure code on exit. |
| michael@0 | 104 | */ |
| michael@0 | 105 | TransliterationRuleData(UErrorCode& status); |
| michael@0 | 106 | |
| michael@0 | 107 | /** |
| michael@0 | 108 | * Copy Constructor |
| michael@0 | 109 | */ |
| michael@0 | 110 | TransliterationRuleData(const TransliterationRuleData&); |
| michael@0 | 111 | |
| michael@0 | 112 | /** |
| michael@0 | 113 | * destructor |
| michael@0 | 114 | */ |
| michael@0 | 115 | ~TransliterationRuleData(); |
| michael@0 | 116 | |
| michael@0 | 117 | /** |
| michael@0 | 118 | * Given a stand-in character, return the UnicodeFunctor that it |
| michael@0 | 119 | * represents, or NULL if it doesn't represent anything. |
| michael@0 | 120 | * @param standIn the given stand-in character. |
| michael@0 | 121 | * @return the UnicodeFunctor that 'standIn' represents |
| michael@0 | 122 | */ |
| michael@0 | 123 | UnicodeFunctor* lookup(UChar32 standIn) const; |
| michael@0 | 124 | |
| michael@0 | 125 | /** |
| michael@0 | 126 | * Given a stand-in character, return the UnicodeMatcher that it |
| michael@0 | 127 | * represents, or NULL if it doesn't represent anything or if it |
| michael@0 | 128 | * represents something that is not a matcher. |
| michael@0 | 129 | * @param standIn the given stand-in character. |
| michael@0 | 130 | * @return return the UnicodeMatcher that 'standIn' represents |
| michael@0 | 131 | */ |
| michael@0 | 132 | UnicodeMatcher* lookupMatcher(UChar32 standIn) const; |
| michael@0 | 133 | |
| michael@0 | 134 | /** |
| michael@0 | 135 | * Given a stand-in character, return the UnicodeReplacer that it |
| michael@0 | 136 | * represents, or NULL if it doesn't represent anything or if it |
| michael@0 | 137 | * represents something that is not a replacer. |
| michael@0 | 138 | * @param standIn the given stand-in character. |
| michael@0 | 139 | * @return return the UnicodeReplacer that 'standIn' represents |
| michael@0 | 140 | */ |
| michael@0 | 141 | UnicodeReplacer* lookupReplacer(UChar32 standIn) const; |
| michael@0 | 142 | |
| michael@0 | 143 | |
| michael@0 | 144 | private: |
| michael@0 | 145 | TransliterationRuleData &operator=(const TransliterationRuleData &other); // forbid copying of this class |
| michael@0 | 146 | }; |
| michael@0 | 147 | |
| michael@0 | 148 | U_NAMESPACE_END |
| michael@0 | 149 | |
| michael@0 | 150 | #endif /* #if !UCONFIG_NO_TRANSLITERATION */ |
| michael@0 | 151 | |
| michael@0 | 152 | #endif |
