projects / apple / icu.git / blame
| Commit | Line | Data |
|---|---|---|
| b75a7d8f A |
1 | /* |
| 2 | * Copyright (C) 1999, International Business Machines Corporation and others. All Rights Reserved. | |
| 3 | ********************************************************************** | |
| 4 | * Date Name Description | |
| 5 | * 11/17/99 aliu Creation. | |
| 6 | ********************************************************************** | |
| 7 | */ | |
| 8 | #ifndef UNIFILT_H | |
| 9 | #define UNIFILT_H | |
| 10 | ||
| 11 | #include "unicode/unifunct.h" | |
| 12 | #include "unicode/unimatch.h" | |
| 13 | ||
| 14 | U_NAMESPACE_BEGIN | |
| 15 | ||
| 16 | /** | |
| 17 | * U_ETHER is used to represent character values for positions outside | |
| 18 | * a range. For example, transliterator uses this to represent | |
| 19 | * characters outside the range contextStart..contextLimit-1. This | |
| 20 | * allows explicit matching by rules and UnicodeSets of text outside a | |
| 21 | * defined range. | |
| 22 | */ | |
| 23 | #define U_ETHER ((UChar)0xFFFF) | |
| 24 | ||
| 25 | /** | |
| 26 | * <code>UnicodeFilter</code> defines a protocol for selecting a | |
| 27 | * subset of the full range (U+0000 to U+10FFFF) of Unicode characters. | |
| 28 | * Currently, filters are used in conjunction with classes like {@link | |
| 29 | * Transliterator} to only process selected characters through a | |
| 30 | * transformation. | |
| 31 | * | |
| 32 | * <p>Note: UnicodeFilter currently stubs out two pure virtual methods | |
| 33 | * of its base class, UnicodeMatcher. These methods are toPattern() | |
| 34 | * and matchesIndexValue(). This is done so that filter classes that | |
| 35 | * are not actually used as matchers -- specifically, those in the | |
| 36 | * UnicodeFilterLogic component, and those in tests -- can continue to | |
| 37 | * work without defining these methods. As long as a filter is not | |
| 38 | * used in an RBT during real transliteration, these methods will not | |
| 39 | * be called. However, this breaks the UnicodeMatcher base class | |
| 40 | * protocol, and it is not a correct solution. | |
| 41 | * | |
| 42 | * <p>In the future we may revisit the UnicodeMatcher / UnicodeFilter | |
| 43 | * hierarchy and either redesign it, or simply remove the stubs in | |
| 44 | * UnicodeFilter and force subclasses to implement the full | |
| 45 | * UnicodeMatcher protocol. | |
| 46 | * | |
| 47 | * @see UnicodeFilterLogic | |
| 48 | * @stable ICU 2.0 | |
| 49 | */ | |
| 50 | class U_COMMON_API UnicodeFilter : public UnicodeFunctor, public UnicodeMatcher { | |
| 51 | ||
| 52 | public: | |
| 53 | /** | |
| 54 | * Destructor | |
| 55 | * @stable ICU 2.0 | |
| 56 | */ | |
| 57 | virtual ~UnicodeFilter(); | |
| 58 | ||
| 59 | /** | |
| 60 | * Returns <tt>true</tt> for characters that are in the selected | |
| 61 | * subset. In other words, if a character is <b>to be | |
| 62 | * filtered</b>, then <tt>contains()</tt> returns | |
| 63 | * <b><tt>false</tt></b>. | |
| 64 | * @stable ICU 2.0 | |
| 65 | */ | |
| 66 | virtual UBool contains(UChar32 c) const = 0; | |
| 67 | ||
| 68 | /** | |
| 69 | * UnicodeFunctor API. Cast 'this' to a UnicodeMatcher* pointer | |
| 70 | * and return the pointer. | |
| 71 | * @draft ICU 2.4 | |
| 72 | */ | |
| 73 | virtual UnicodeMatcher* toMatcher() const; | |
| 74 | ||
| 75 | /** | |
| 76 | * Implement UnicodeMatcher API. | |
| 77 | * @draft ICU 2.4 | |
| 78 | */ | |
| 79 | virtual UMatchDegree matches(const Replaceable& text, | |
| 80 | int32_t& offset, | |
| 81 | int32_t limit, | |
| 82 | UBool incremental); | |
| 83 | ||
| 84 | /** | |
| 85 | * UnicodeFunctor API. Nothing to do. | |
| 86 | * @draft ICU 2.4 | |
| 87 | */ | |
| 88 | virtual void setData(const TransliterationRuleData*) {} | |
| 89 | ||
| 90 | /** | |
| 91 | * ICU "poor man's RTTI", returns a UClassID for the actual class. | |
| 92 | * | |
| 93 | * @draft ICU 2.2 | |
| 94 | */ | |
| 95 | virtual inline UClassID getDynamicClassID() const = 0; | |
| 96 | ||
| 97 | /** | |
| 98 | * ICU "poor man's RTTI", returns a UClassID for this class. | |
| 99 | * | |
| 100 | * @draft ICU 2.2 | |
| 101 | */ | |
| 102 | static inline UClassID getStaticClassID() { return (UClassID)&fgClassID; } | |
| 103 | ||
| 104 | protected: | |
| 105 | ||
| 106 | /** | |
| 107 | * @stable ICU 2.0 | |
| 108 | */ | |
| 109 | UnicodeFilter(); | |
| 110 | ||
| 111 | private: | |
| 112 | ||
| 113 | /** | |
| 114 | * The address of this static class variable serves as this class's ID | |
| 115 | * for ICU "poor man's RTTI". | |
| 116 | */ | |
| 117 | static const char fgClassID; | |
| 118 | }; | |
| 119 | ||
| 120 | inline UnicodeFilter::UnicodeFilter() {} | |
| 121 | inline UnicodeFilter::~UnicodeFilter() {} | |
| 122 | ||
| 123 | U_NAMESPACE_END | |
| 124 | ||
| 125 | #endif |