icuSources/common/unicode/unifilt.h

   1 /*
   2 * Copyright (C) 1999-2004, International Business Machines Corporation and others.
   3 * All Rights Reserved.
   4 **********************************************************************
   5 *   Date        Name        Description
   6 *   11/17/99    aliu        Creation.
   7 **********************************************************************
   8 */
   9 #ifndef UNIFILT_H
  10 #define UNIFILT_H
  11
  12 #include "unicode/unifunct.h"
  13 #include "unicode/unimatch.h"
  14
  15 U_NAMESPACE_BEGIN
  16
  17 /**
  18  * U_ETHER is used to represent character values for positions outside
  19  * a range.  For example, transliterator uses this to represent
  20  * characters outside the range contextStart..contextLimit-1.  This
  21  * allows explicit matching by rules and UnicodeSets of text outside a
  22  * defined range.
  23  * @draft ICU 3.0
  24  */
  25 #define U_ETHER ((UChar)0xFFFF)
  26
  27 /**
  28  * <code>UnicodeFilter</code> defines a protocol for selecting a
  29  * subset of the full range (U+0000 to U+10FFFF) of Unicode characters.
  30  * Currently, filters are used in conjunction with classes like {@link
  31  * Transliterator} to only process selected characters through a
  32  * transformation.
  33  *
  34  * <p>Note: UnicodeFilter currently stubs out two pure virtual methods
  35  * of its base class, UnicodeMatcher.  These methods are toPattern()
  36  * and matchesIndexValue().  This is done so that filter classes that
  37  * are not actually used as matchers -- specifically, those in the
  38  * UnicodeFilterLogic component, and those in tests -- can continue to
  39  * work without defining these methods.  As long as a filter is not
  40  * used in an RBT during real transliteration, these methods will not
  41  * be called.  However, this breaks the UnicodeMatcher base class
  42  * protocol, and it is not a correct solution.
  43  *
  44  * <p>In the future we may revisit the UnicodeMatcher / UnicodeFilter
  45  * hierarchy and either redesign it, or simply remove the stubs in
  46  * UnicodeFilter and force subclasses to implement the full
  47  * UnicodeMatcher protocol.
  48  *
  49  * @see UnicodeFilterLogic
  50  * @stable ICU 2.0
  51  */
  52 class U_COMMON_API UnicodeFilter : public UnicodeFunctor, public UnicodeMatcher {
  53
  54 public:
  55     /**
  56      * Destructor
  57      * @stable ICU 2.0
  58      */
  59     virtual ~UnicodeFilter();
  60
  61     /**
  62      * Returns <tt>true</tt> for characters that are in the selected
  63      * subset.  In other words, if a character is <b>to be
  64      * filtered</b>, then <tt>contains()</tt> returns
  65      * <b><tt>false</tt></b>.
  66      * @stable ICU 2.0
  67      */
  68     virtual UBool contains(UChar32 c) const = 0;
  69
  70     /**
  71      * UnicodeFunctor API.  Cast 'this' to a UnicodeMatcher* pointer
  72      * and return the pointer.
  73      * @stable ICU 2.4
  74      */
  75     virtual UnicodeMatcher* toMatcher() const;
  76
  77     /**
  78      * Implement UnicodeMatcher API.
  79      * @stable ICU 2.4
  80      */
  81     virtual UMatchDegree matches(const Replaceable& text,
  82                                  int32_t& offset,
  83                                  int32_t limit,
  84                                  UBool incremental);
  85
  86     /**
  87      * UnicodeFunctor API.  Nothing to do.
  88      * @stable ICU 2.4
  89      */
  90     virtual void setData(const TransliterationRuleData*);
  91
  92     /**
  93      * ICU "poor man's RTTI", returns a UClassID for the actual class.
  94      *
  95      * @stable ICU 2.2
  96      */
  97     virtual UClassID getDynamicClassID() const = 0;
  98
  99     /**
 100      * ICU "poor man's RTTI", returns a UClassID for this class.
 101      *
 102      * @stable ICU 2.2
 103      */
 104     static UClassID U_EXPORT2 getStaticClassID();
 105
 106 protected:
 107
 108     /*
 109      * Since this class has pure virtual functions,
 110      * a constructor can't be used.
 111      * @stable ICU 2.0
 112      */
 113 /*    UnicodeFilter();*/
 114 };
 115
 116 /*inline UnicodeFilter::UnicodeFilter() {}*/
 117
 118 U_NAMESPACE_END
 119
 120 #endif