2 **************************************************************************
3 * Copyright (C) 1999-2005, International Business Machines Corporation and
4 * others. All Rights Reserved.
5 **************************************************************************
6 * Date Name Description
7 * 11/17/99 aliu Creation. Ported from java. Modified to
8 * match current UnicodeString API. Forced
9 * to use name "handleReplaceBetween" because
10 * of existing methods in UnicodeString.
11 **************************************************************************
17 #include "unicode/uobject.h"
21 * \brief C++ API: Replaceable String
29 * <code>Replaceable</code> is an abstract base class representing a
30 * string of characters that supports the replacement of a range of
31 * itself with a new string of characters. It is used by APIs that
32 * change a piece of text while retaining metadata. Metadata is data
33 * other than the Unicode characters returned by char32At(). One
34 * example of metadata is style attributes; another is an edit
35 * history, marking each character with an author and revision number.
37 * <p>An implicit aspect of the <code>Replaceable</code> API is that
38 * during a replace operation, new characters take on the metadata of
39 * the old characters. For example, if the string "the <b>bold</b>
40 * font" has range (4, 8) replaced with "strong", then it becomes "the
41 * <b>strong</b> font".
43 * <p><code>Replaceable</code> specifies ranges using a start
44 * offset and a limit offset. The range of characters thus specified
45 * includes the characters at offset start..limit-1. That is, the
46 * start offset is inclusive, and the limit offset is exclusive.
48 * <p><code>Replaceable</code> also includes API to access characters
49 * in the string: <code>length()</code>, <code>charAt()</code>,
50 * <code>char32At()</code>, and <code>extractBetween()</code>.
52 * <p>For a subclass to support metadata, typical behavior of
53 * <code>replace()</code> is the following:
55 * <li>Set the metadata of the new text to the metadata of the first
56 * character replaced</li>
57 * <li>If no characters are replaced, use the metadata of the
58 * previous character</li>
59 * <li>If there is no previous character (i.e. start == 0), use the
60 * following character</li>
61 * <li>If there is no following character (i.e. the replaceable was
62 * empty), use default metadata.<br>
63 * <li>If the code point U+FFFF is seen, it should be interpreted as
64 * a special marker having no metadata<li>
67 * If this is not the behavior, the subclass should document any differences.
71 class U_COMMON_API Replaceable : public UObject {
78 virtual ~Replaceable();
81 * Returns the number of 16-bit code units in the text.
82 * @return number of 16-bit code units in text
85 inline int32_t length() const;
88 * Returns the 16-bit code unit at the given offset into the text.
89 * @param offset an integer between 0 and <code>length()</code>-1
91 * @return 16-bit code unit of text at given offset
94 inline UChar charAt(int32_t offset) const;
97 * Returns the 32-bit code point at the given 16-bit offset into
98 * the text. This assumes the text is stored as 16-bit code units
99 * with surrogate pairs intermixed. If the offset of a leading or
100 * trailing code unit of a surrogate pair is given, return the
101 * code point of the surrogate pair.
103 * @param offset an integer between 0 and <code>length()</code>-1
105 * @return 32-bit code point of text at given offset
108 inline UChar32 char32At(int32_t offset) const;
111 * Copies characters in the range [<tt>start</tt>, <tt>limit</tt>)
112 * into the UnicodeString <tt>target</tt>.
113 * @param start offset of first character which will be copied
114 * @param limit offset immediately following the last character to
116 * @param target UnicodeString into which to copy characters.
117 * @return A reference to <TT>target</TT>
120 virtual void extractBetween(int32_t start,
122 UnicodeString& target) const = 0;
125 * Replaces a substring of this object with the given text. If the
126 * characters being replaced have metadata, the new characters
127 * that replace them should be given the same metadata.
129 * <p>Subclasses must ensure that if the text between start and
130 * limit is equal to the replacement text, that replace has no
131 * effect. That is, any metadata
132 * should be unaffected. In addition, subclasses are encouraged to
133 * check for initial and trailing identical characters, and make a
134 * smaller replacement if possible. This will preserve as much
135 * metadata as possible.
136 * @param start the beginning index, inclusive; <code>0 <= start
138 * @param limit the ending index, exclusive; <code>start <= limit
139 * <= length()</code>.
140 * @param text the text to replace characters <code>start</code>
141 * to <code>limit - 1</code>
144 virtual void handleReplaceBetween(int32_t start,
146 const UnicodeString& text) = 0;
147 // Note: All other methods in this class take the names of
148 // existing UnicodeString methods. This method is the exception.
149 // It is named differently because all replace methods of
150 // UnicodeString return a UnicodeString&. The 'between' is
151 // required in order to conform to the UnicodeString naming
152 // convention; API taking start/length are named <operation>, and
153 // those taking start/limit are named <operationBetween>. The
154 // 'handle' is added because 'replaceBetween' and
155 // 'doReplaceBetween' are already taken.
158 * Copies a substring of this object, retaining metadata.
159 * This method is used to duplicate or reorder substrings.
160 * The destination index must not overlap the source range.
162 * @param start the beginning index, inclusive; <code>0 <= start <=
164 * @param limit the ending index, exclusive; <code>start <= limit <=
166 * @param dest the destination index. The characters from
167 * <code>start..limit-1</code> will be copied to <code>dest</code>.
168 * Implementations of this method may assume that <code>dest <= start ||
169 * dest >= limit</code>.
172 virtual void copy(int32_t start, int32_t limit, int32_t dest) = 0;
175 * Returns true if this object contains metadata. If a
176 * Replaceable object has metadata, calls to the Replaceable API
177 * must be made so as to preserve metadata. If it does not, calls
178 * to the Replaceable API may be optimized to improve performance.
179 * The default implementation returns true.
180 * @return true if this object contains metadata
183 virtual UBool hasMetaData() const;
186 * Clone this object, an instance of a subclass of Replaceable.
187 * Clones can be used concurrently in multiple threads.
188 * If a subclass does not implement clone(), or if an error occurs,
189 * then NULL is returned.
190 * The clone functions in all subclasses return a pointer to a Replaceable
191 * because some compilers do not support covariant (same-as-this)
192 * return types; cast to the appropriate subclass if necessary.
193 * The caller must delete the clone.
195 * @return a clone of this object
197 * @see getDynamicClassID
200 virtual Replaceable *clone() const;
205 * Default constructor.
211 * Assignment operator not declared. The compiler will provide one
212 * which does nothing since this class does not contain any data members.
213 * API/code coverage may show the assignment operator as present and
215 * Subclasses need this assignment operator if they use compiler-provided
216 * assignment operators of their own. An alternative to not declaring one
217 * here would be to declare and empty-implement a protected or public one.
218 Replaceable &Replaceable::operator=(const Replaceable &);
222 * Virtual version of length().
225 virtual int32_t getLength() const = 0;
228 * Virtual version of charAt().
231 virtual UChar getCharAt(int32_t offset) const = 0;
234 * Virtual version of char32At().
237 virtual UChar32 getChar32At(int32_t offset) const = 0;
241 Replaceable::length() const {
246 Replaceable::charAt(int32_t offset) const {
247 return getCharAt(offset);
251 Replaceable::char32At(int32_t offset) const {
252 return getChar32At(offset);
255 // There is no rep.cpp, see unistr.cpp for Replaceable function implementations.