1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/common/unicode/unifilt.h Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,120 @@
1.4 +/*
1.5 +**********************************************************************
1.6 +* Copyright (C) 1999-2010, International Business Machines Corporation and others.
1.7 +* All Rights Reserved.
1.8 +**********************************************************************
1.9 +* Date Name Description
1.10 +* 11/17/99 aliu Creation.
1.11 +**********************************************************************
1.12 +*/
1.13 +#ifndef UNIFILT_H
1.14 +#define UNIFILT_H
1.15 +
1.16 +#include "unicode/unifunct.h"
1.17 +#include "unicode/unimatch.h"
1.18 +
1.19 +/**
1.20 + * \file
1.21 + * \brief C++ API: Unicode Filter
1.22 + */
1.23 +
1.24 +U_NAMESPACE_BEGIN
1.25 +
1.26 +/**
1.27 + * U_ETHER is used to represent character values for positions outside
1.28 + * a range. For example, transliterator uses this to represent
1.29 + * characters outside the range contextStart..contextLimit-1. This
1.30 + * allows explicit matching by rules and UnicodeSets of text outside a
1.31 + * defined range.
1.32 + * @stable ICU 3.0
1.33 + */
1.34 +#define U_ETHER ((UChar)0xFFFF)
1.35 +
1.36 +/**
1.37 + *
1.38 + * <code>UnicodeFilter</code> defines a protocol for selecting a
1.39 + * subset of the full range (U+0000 to U+10FFFF) of Unicode characters.
1.40 + * Currently, filters are used in conjunction with classes like {@link
1.41 + * Transliterator} to only process selected characters through a
1.42 + * transformation.
1.43 + *
1.44 + * <p>Note: UnicodeFilter currently stubs out two pure virtual methods
1.45 + * of its base class, UnicodeMatcher. These methods are toPattern()
1.46 + * and matchesIndexValue(). This is done so that filter classes that
1.47 + * are not actually used as matchers -- specifically, those in the
1.48 + * UnicodeFilterLogic component, and those in tests -- can continue to
1.49 + * work without defining these methods. As long as a filter is not
1.50 + * used in an RBT during real transliteration, these methods will not
1.51 + * be called. However, this breaks the UnicodeMatcher base class
1.52 + * protocol, and it is not a correct solution.
1.53 + *
1.54 + * <p>In the future we may revisit the UnicodeMatcher / UnicodeFilter
1.55 + * hierarchy and either redesign it, or simply remove the stubs in
1.56 + * UnicodeFilter and force subclasses to implement the full
1.57 + * UnicodeMatcher protocol.
1.58 + *
1.59 + * @see UnicodeFilterLogic
1.60 + * @stable ICU 2.0
1.61 + */
1.62 +class U_COMMON_API UnicodeFilter : public UnicodeFunctor, public UnicodeMatcher {
1.63 +
1.64 +public:
1.65 + /**
1.66 + * Destructor
1.67 + * @stable ICU 2.0
1.68 + */
1.69 + virtual ~UnicodeFilter();
1.70 +
1.71 + /**
1.72 + * Returns <tt>true</tt> for characters that are in the selected
1.73 + * subset. In other words, if a character is <b>to be
1.74 + * filtered</b>, then <tt>contains()</tt> returns
1.75 + * <b><tt>false</tt></b>.
1.76 + * @stable ICU 2.0
1.77 + */
1.78 + virtual UBool contains(UChar32 c) const = 0;
1.79 +
1.80 + /**
1.81 + * UnicodeFunctor API. Cast 'this' to a UnicodeMatcher* pointer
1.82 + * and return the pointer.
1.83 + * @stable ICU 2.4
1.84 + */
1.85 + virtual UnicodeMatcher* toMatcher() const;
1.86 +
1.87 + /**
1.88 + * Implement UnicodeMatcher API.
1.89 + * @stable ICU 2.4
1.90 + */
1.91 + virtual UMatchDegree matches(const Replaceable& text,
1.92 + int32_t& offset,
1.93 + int32_t limit,
1.94 + UBool incremental);
1.95 +
1.96 + /**
1.97 + * UnicodeFunctor API. Nothing to do.
1.98 + * @stable ICU 2.4
1.99 + */
1.100 + virtual void setData(const TransliterationRuleData*);
1.101 +
1.102 + /**
1.103 + * ICU "poor man's RTTI", returns a UClassID for this class.
1.104 + *
1.105 + * @stable ICU 2.2
1.106 + */
1.107 + static UClassID U_EXPORT2 getStaticClassID();
1.108 +
1.109 +protected:
1.110 +
1.111 + /*
1.112 + * Since this class has pure virtual functions,
1.113 + * a constructor can't be used.
1.114 + * @stable ICU 2.0
1.115 + */
1.116 +/* UnicodeFilter();*/
1.117 +};
1.118 +
1.119 +/*inline UnicodeFilter::UnicodeFilter() {}*/
1.120 +
1.121 +U_NAMESPACE_END
1.122 +
1.123 +#endif
