jpayne@69: // © 2016 and later: Unicode, Inc. and others.
jpayne@69: // License & terms of use: http://www.unicode.org/copyright.html
jpayne@69: /*
jpayne@69: ********************************************************************
jpayne@69: *
jpayne@69: *   Copyright (C) 1997-2011, International Business Machines
jpayne@69: *   Corporation and others.  All Rights Reserved.
jpayne@69: *
jpayne@69: ********************************************************************
jpayne@69: */
jpayne@69: 
jpayne@69: #ifndef CHARITER_H
jpayne@69: #define CHARITER_H
jpayne@69: 
jpayne@69: #include "unicode/utypes.h"
jpayne@69: 
jpayne@69: #if U_SHOW_CPLUSPLUS_API
jpayne@69: 
jpayne@69: #include "unicode/uobject.h"
jpayne@69: #include "unicode/unistr.h"
jpayne@69: /**
jpayne@69:  * \file
jpayne@69:  * \brief C++ API: Character Iterator
jpayne@69:  */
jpayne@69:  
jpayne@69: U_NAMESPACE_BEGIN
jpayne@69: /**
jpayne@69:  * Abstract class that defines an API for forward-only iteration
jpayne@69:  * on text objects.
jpayne@69:  * This is a minimal interface for iteration without random access
jpayne@69:  * or backwards iteration. It is especially useful for wrapping
jpayne@69:  * streams with converters into an object for collation or
jpayne@69:  * normalization.
jpayne@69:  *
jpayne@69:  * <p>Characters can be accessed in two ways: as code units or as
jpayne@69:  * code points.
jpayne@69:  * Unicode code points are 21-bit integers and are the scalar values
jpayne@69:  * of Unicode characters. ICU uses the type UChar32 for them.
jpayne@69:  * Unicode code units are the storage units of a given
jpayne@69:  * Unicode/UCS Transformation Format (a character encoding scheme).
jpayne@69:  * With UTF-16, all code points can be represented with either one
jpayne@69:  * or two code units ("surrogates").
jpayne@69:  * String storage is typically based on code units, while properties
jpayne@69:  * of characters are typically determined using code point values.
jpayne@69:  * Some processes may be designed to work with sequences of code units,
jpayne@69:  * or it may be known that all characters that are important to an
jpayne@69:  * algorithm can be represented with single code units.
jpayne@69:  * Other processes will need to use the code point access functions.</p>
jpayne@69:  *
jpayne@69:  * <p>ForwardCharacterIterator provides nextPostInc() to access
jpayne@69:  * a code unit and advance an internal position into the text object,
jpayne@69:  * similar to a <code>return text[position++]</code>.<br>
jpayne@69:  * It provides next32PostInc() to access a code point and advance an internal
jpayne@69:  * position.</p>
jpayne@69:  *
jpayne@69:  * <p>next32PostInc() assumes that the current position is that of
jpayne@69:  * the beginning of a code point, i.e., of its first code unit.
jpayne@69:  * After next32PostInc(), this will be true again.
jpayne@69:  * In general, access to code units and code points in the same
jpayne@69:  * iteration loop should not be mixed. In UTF-16, if the current position
jpayne@69:  * is on a second code unit (Low Surrogate), then only that code unit
jpayne@69:  * is returned even by next32PostInc().</p>
jpayne@69:  *
jpayne@69:  * <p>For iteration with either function, there are two ways to
jpayne@69:  * check for the end of the iteration. When there are no more
jpayne@69:  * characters in the text object:
jpayne@69:  * <ul>
jpayne@69:  * <li>The hasNext() function returns FALSE.</li>
jpayne@69:  * <li>nextPostInc() and next32PostInc() return DONE
jpayne@69:  *     when one attempts to read beyond the end of the text object.</li>
jpayne@69:  * </ul>
jpayne@69:  *
jpayne@69:  * Example:
jpayne@69:  * \code 
jpayne@69:  * void function1(ForwardCharacterIterator &it) {
jpayne@69:  *     UChar32 c;
jpayne@69:  *     while(it.hasNext()) {
jpayne@69:  *         c=it.next32PostInc();
jpayne@69:  *         // use c
jpayne@69:  *     }
jpayne@69:  * }
jpayne@69:  *
jpayne@69:  * void function1(ForwardCharacterIterator &it) {
jpayne@69:  *     char16_t c;
jpayne@69:  *     while((c=it.nextPostInc())!=ForwardCharacterIterator::DONE) {
jpayne@69:  *         // use c
jpayne@69:  *      }
jpayne@69:  *  }
jpayne@69:  * \endcode
jpayne@69:  * </p>
jpayne@69:  *
jpayne@69:  * @stable ICU 2.0
jpayne@69:  */
jpayne@69: class U_COMMON_API ForwardCharacterIterator : public UObject {
jpayne@69: public:
jpayne@69:     /**
jpayne@69:      * Value returned by most of ForwardCharacterIterator's functions
jpayne@69:      * when the iterator has reached the limits of its iteration.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     enum { DONE = 0xffff };
jpayne@69:     
jpayne@69:     /**
jpayne@69:      * Destructor.  
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual ~ForwardCharacterIterator();
jpayne@69:     
jpayne@69:     /**
jpayne@69:      * Returns true when both iterators refer to the same
jpayne@69:      * character in the same character-storage object.  
jpayne@69:      * @param that The ForwardCharacterIterator to be compared for equality
jpayne@69:      * @return true when both iterators refer to the same
jpayne@69:      * character in the same character-storage object
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual UBool operator==(const ForwardCharacterIterator& that) const = 0;
jpayne@69:     
jpayne@69:     /**
jpayne@69:      * Returns true when the iterators refer to different
jpayne@69:      * text-storage objects, or to different characters in the
jpayne@69:      * same text-storage object.  
jpayne@69:      * @param that The ForwardCharacterIterator to be compared for inequality
jpayne@69:      * @return true when the iterators refer to different
jpayne@69:      * text-storage objects, or to different characters in the
jpayne@69:      * same text-storage object
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     inline UBool operator!=(const ForwardCharacterIterator& that) const;
jpayne@69:     
jpayne@69:     /**
jpayne@69:      * Generates a hash code for this iterator.  
jpayne@69:      * @return the hash code.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual int32_t hashCode(void) const = 0;
jpayne@69:     
jpayne@69:     /**
jpayne@69:      * Returns a UClassID for this ForwardCharacterIterator ("poor man's
jpayne@69:      * RTTI").<P> Despite the fact that this function is public,
jpayne@69:      * DO NOT CONSIDER IT PART OF CHARACTERITERATOR'S API! 
jpayne@69:      * @return a UClassID for this ForwardCharacterIterator 
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual UClassID getDynamicClassID(void) const = 0;
jpayne@69:     
jpayne@69:     /**
jpayne@69:      * Gets the current code unit for returning and advances to the next code unit
jpayne@69:      * in the iteration range
jpayne@69:      * (toward endIndex()).  If there are
jpayne@69:      * no more code units to return, returns DONE.
jpayne@69:      * @return the current code unit.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual char16_t         nextPostInc(void) = 0;
jpayne@69:     
jpayne@69:     /**
jpayne@69:      * Gets the current code point for returning and advances to the next code point
jpayne@69:      * in the iteration range
jpayne@69:      * (toward endIndex()).  If there are
jpayne@69:      * no more code points to return, returns DONE.
jpayne@69:      * @return the current code point.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual UChar32       next32PostInc(void) = 0;
jpayne@69:     
jpayne@69:     /**
jpayne@69:      * Returns FALSE if there are no more code units or code points
jpayne@69:      * at or after the current position in the iteration range.
jpayne@69:      * This is used with nextPostInc() or next32PostInc() in forward
jpayne@69:      * iteration.
jpayne@69:      * @returns FALSE if there are no more code units or code points
jpayne@69:      * at or after the current position in the iteration range.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual UBool        hasNext() = 0;
jpayne@69:     
jpayne@69: protected:
jpayne@69:     /** Default constructor to be overridden in the implementing class. @stable ICU 2.0*/
jpayne@69:     ForwardCharacterIterator();
jpayne@69:     
jpayne@69:     /** Copy constructor to be overridden in the implementing class. @stable ICU 2.0*/
jpayne@69:     ForwardCharacterIterator(const ForwardCharacterIterator &other);
jpayne@69:     
jpayne@69:     /**
jpayne@69:      * Assignment operator to be overridden in the implementing class.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     ForwardCharacterIterator &operator=(const ForwardCharacterIterator&) { return *this; }
jpayne@69: };
jpayne@69: 
jpayne@69: /**
jpayne@69:  * Abstract class that defines an API for iteration
jpayne@69:  * on text objects.
jpayne@69:  * This is an interface for forward and backward iteration
jpayne@69:  * and random access into a text object.
jpayne@69:  *
jpayne@69:  * <p>The API provides backward compatibility to the Java and older ICU
jpayne@69:  * CharacterIterator classes but extends them significantly:
jpayne@69:  * <ol>
jpayne@69:  * <li>CharacterIterator is now a subclass of ForwardCharacterIterator.</li>
jpayne@69:  * <li>While the old API functions provided forward iteration with
jpayne@69:  *     "pre-increment" semantics, the new one also provides functions
jpayne@69:  *     with "post-increment" semantics. They are more efficient and should
jpayne@69:  *     be the preferred iterator functions for new implementations.
jpayne@69:  *     The backward iteration always had "pre-decrement" semantics, which
jpayne@69:  *     are efficient.</li>
jpayne@69:  * <li>Just like ForwardCharacterIterator, it provides access to
jpayne@69:  *     both code units and code points. Code point access versions are available
jpayne@69:  *     for the old and the new iteration semantics.</li>
jpayne@69:  * <li>There are new functions for setting and moving the current position
jpayne@69:  *     without returning a character, for efficiency.</li>
jpayne@69:  * </ol>
jpayne@69:  *
jpayne@69:  * See ForwardCharacterIterator for examples for using the new forward iteration
jpayne@69:  * functions. For backward iteration, there is also a hasPrevious() function
jpayne@69:  * that can be used analogously to hasNext().
jpayne@69:  * The old functions work as before and are shown below.</p>
jpayne@69:  *
jpayne@69:  * <p>Examples for some of the new functions:</p>
jpayne@69:  *
jpayne@69:  * Forward iteration with hasNext():
jpayne@69:  * \code
jpayne@69:  * void forward1(CharacterIterator &it) {
jpayne@69:  *     UChar32 c;
jpayne@69:  *     for(it.setToStart(); it.hasNext();) {
jpayne@69:  *         c=it.next32PostInc();
jpayne@69:  *         // use c
jpayne@69:  *     }
jpayne@69:  *  }
jpayne@69:  * \endcode
jpayne@69:  * Forward iteration more similar to loops with the old forward iteration,
jpayne@69:  * showing a way to convert simple for() loops:
jpayne@69:  * \code
jpayne@69:  * void forward2(CharacterIterator &it) {
jpayne@69:  *     char16_t c;
jpayne@69:  *     for(c=it.firstPostInc(); c!=CharacterIterator::DONE; c=it.nextPostInc()) {
jpayne@69:  *          // use c
jpayne@69:  *      }
jpayne@69:  * }
jpayne@69:  * \endcode
jpayne@69:  * Backward iteration with setToEnd() and hasPrevious():
jpayne@69:  * \code
jpayne@69:  *  void backward1(CharacterIterator &it) {
jpayne@69:  *      UChar32 c;
jpayne@69:  *      for(it.setToEnd(); it.hasPrevious();) {
jpayne@69:  *         c=it.previous32();
jpayne@69:  *          // use c
jpayne@69:  *      }
jpayne@69:  *  }
jpayne@69:  * \endcode
jpayne@69:  * Backward iteration with a more traditional for() loop:
jpayne@69:  * \code
jpayne@69:  * void backward2(CharacterIterator &it) {
jpayne@69:  *     char16_t c;
jpayne@69:  *     for(c=it.last(); c!=CharacterIterator::DONE; c=it.previous()) {
jpayne@69:  *         // use c
jpayne@69:  *      }
jpayne@69:  *  }
jpayne@69:  * \endcode
jpayne@69:  *
jpayne@69:  * Example for random access:
jpayne@69:  * \code
jpayne@69:  *  void random(CharacterIterator &it) {
jpayne@69:  *      // set to the third code point from the beginning
jpayne@69:  *      it.move32(3, CharacterIterator::kStart);
jpayne@69:  *      // get a code point from here without moving the position
jpayne@69:  *      UChar32 c=it.current32();
jpayne@69:  *      // get the position
jpayne@69:  *      int32_t pos=it.getIndex();
jpayne@69:  *      // get the previous code unit
jpayne@69:  *      char16_t u=it.previous();
jpayne@69:  *      // move back one more code unit
jpayne@69:  *      it.move(-1, CharacterIterator::kCurrent);
jpayne@69:  *      // set the position back to where it was
jpayne@69:  *      // and read the same code point c and move beyond it
jpayne@69:  *      it.setIndex(pos);
jpayne@69:  *      if(c!=it.next32PostInc()) {
jpayne@69:  *          exit(1); // CharacterIterator inconsistent
jpayne@69:  *      }
jpayne@69:  *  }
jpayne@69:  * \endcode
jpayne@69:  *
jpayne@69:  * <p>Examples, especially for the old API:</p>
jpayne@69:  *
jpayne@69:  * Function processing characters, in this example simple output
jpayne@69:  * <pre>
jpayne@69:  * \code
jpayne@69:  *  void processChar( char16_t c )
jpayne@69:  *  {
jpayne@69:  *      cout << " " << c;
jpayne@69:  *  }
jpayne@69:  * \endcode
jpayne@69:  * </pre>
jpayne@69:  * Traverse the text from start to finish
jpayne@69:  * <pre> 
jpayne@69:  * \code
jpayne@69:  *  void traverseForward(CharacterIterator& iter)
jpayne@69:  *  {
jpayne@69:  *      for(char16_t c = iter.first(); c != CharacterIterator.DONE; c = iter.next()) {
jpayne@69:  *          processChar(c);
jpayne@69:  *      }
jpayne@69:  *  }
jpayne@69:  * \endcode
jpayne@69:  * </pre>
jpayne@69:  * Traverse the text backwards, from end to start
jpayne@69:  * <pre>
jpayne@69:  * \code
jpayne@69:  *  void traverseBackward(CharacterIterator& iter)
jpayne@69:  *  {
jpayne@69:  *      for(char16_t c = iter.last(); c != CharacterIterator.DONE; c = iter.previous()) {
jpayne@69:  *          processChar(c);
jpayne@69:  *      }
jpayne@69:  *  }
jpayne@69:  * \endcode
jpayne@69:  * </pre>
jpayne@69:  * Traverse both forward and backward from a given position in the text. 
jpayne@69:  * Calls to notBoundary() in this example represents some additional stopping criteria.
jpayne@69:  * <pre>
jpayne@69:  * \code
jpayne@69:  * void traverseOut(CharacterIterator& iter, int32_t pos)
jpayne@69:  * {
jpayne@69:  *      char16_t c;
jpayne@69:  *      for (c = iter.setIndex(pos);
jpayne@69:  *      c != CharacterIterator.DONE && (Unicode::isLetter(c) || Unicode::isDigit(c));
jpayne@69:  *          c = iter.next()) {}
jpayne@69:  *      int32_t end = iter.getIndex();
jpayne@69:  *      for (c = iter.setIndex(pos);
jpayne@69:  *          c != CharacterIterator.DONE && (Unicode::isLetter(c) || Unicode::isDigit(c));
jpayne@69:  *          c = iter.previous()) {}
jpayne@69:  *      int32_t start = iter.getIndex() + 1;
jpayne@69:  *  
jpayne@69:  *      cout << "start: " << start << " end: " << end << endl;
jpayne@69:  *      for (c = iter.setIndex(start); iter.getIndex() < end; c = iter.next() ) {
jpayne@69:  *          processChar(c);
jpayne@69:  *     }
jpayne@69:  *  }
jpayne@69:  * \endcode
jpayne@69:  * </pre>
jpayne@69:  * Creating a StringCharacterIterator and calling the test functions
jpayne@69:  * <pre>
jpayne@69:  * \code
jpayne@69:  *  void CharacterIterator_Example( void )
jpayne@69:  *   {
jpayne@69:  *       cout << endl << "===== CharacterIterator_Example: =====" << endl;
jpayne@69:  *       UnicodeString text("Ein kleiner Satz.");
jpayne@69:  *       StringCharacterIterator iterator(text);
jpayne@69:  *       cout << "----- traverseForward: -----------" << endl;
jpayne@69:  *       traverseForward( iterator );
jpayne@69:  *       cout << endl << endl << "----- traverseBackward: ----------" << endl;
jpayne@69:  *       traverseBackward( iterator );
jpayne@69:  *       cout << endl << endl << "----- traverseOut: ---------------" << endl;
jpayne@69:  *       traverseOut( iterator, 7 );
jpayne@69:  *       cout << endl << endl << "-----" << endl;
jpayne@69:  *   }
jpayne@69:  * \endcode
jpayne@69:  * </pre>
jpayne@69:  *
jpayne@69:  * @stable ICU 2.0
jpayne@69:  */
jpayne@69: class U_COMMON_API CharacterIterator : public ForwardCharacterIterator {
jpayne@69: public:
jpayne@69:     /**
jpayne@69:      * Origin enumeration for the move() and move32() functions.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     enum EOrigin { kStart, kCurrent, kEnd };
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Destructor.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual ~CharacterIterator();
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Returns a pointer to a new CharacterIterator of the same
jpayne@69:      * concrete class as this one, and referring to the same
jpayne@69:      * character in the same text-storage object as this one.  The
jpayne@69:      * caller is responsible for deleting the new clone.  
jpayne@69:      * @return a pointer to a new CharacterIterator
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual CharacterIterator* clone() const = 0;
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Sets the iterator to refer to the first code unit in its
jpayne@69:      * iteration range, and returns that code unit.
jpayne@69:      * This can be used to begin an iteration with next().
jpayne@69:      * @return the first code unit in its iteration range.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual char16_t         first(void) = 0;
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Sets the iterator to refer to the first code unit in its
jpayne@69:      * iteration range, returns that code unit, and moves the position
jpayne@69:      * to the second code unit. This is an alternative to setToStart()
jpayne@69:      * for forward iteration with nextPostInc().
jpayne@69:      * @return the first code unit in its iteration range.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual char16_t         firstPostInc(void);
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Sets the iterator to refer to the first code point in its
jpayne@69:      * iteration range, and returns that code unit,
jpayne@69:      * This can be used to begin an iteration with next32().
jpayne@69:      * Note that an iteration with next32PostInc(), beginning with,
jpayne@69:      * e.g., setToStart() or firstPostInc(), is more efficient.
jpayne@69:      * @return the first code point in its iteration range.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual UChar32       first32(void) = 0;
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Sets the iterator to refer to the first code point in its
jpayne@69:      * iteration range, returns that code point, and moves the position
jpayne@69:      * to the second code point. This is an alternative to setToStart()
jpayne@69:      * for forward iteration with next32PostInc().
jpayne@69:      * @return the first code point in its iteration range.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual UChar32       first32PostInc(void);
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Sets the iterator to refer to the first code unit or code point in its
jpayne@69:      * iteration range. This can be used to begin a forward
jpayne@69:      * iteration with nextPostInc() or next32PostInc().
jpayne@69:      * @return the start position of the iteration range
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     inline int32_t    setToStart();
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Sets the iterator to refer to the last code unit in its
jpayne@69:      * iteration range, and returns that code unit.
jpayne@69:      * This can be used to begin an iteration with previous().
jpayne@69:      * @return the last code unit.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual char16_t         last(void) = 0;
jpayne@69:         
jpayne@69:     /**
jpayne@69:      * Sets the iterator to refer to the last code point in its
jpayne@69:      * iteration range, and returns that code unit.
jpayne@69:      * This can be used to begin an iteration with previous32().
jpayne@69:      * @return the last code point.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual UChar32       last32(void) = 0;
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Sets the iterator to the end of its iteration range, just behind
jpayne@69:      * the last code unit or code point. This can be used to begin a backward
jpayne@69:      * iteration with previous() or previous32().
jpayne@69:      * @return the end position of the iteration range
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     inline int32_t    setToEnd();
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Sets the iterator to refer to the "position"-th code unit
jpayne@69:      * in the text-storage object the iterator refers to, and
jpayne@69:      * returns that code unit.  
jpayne@69:      * @param position the "position"-th code unit in the text-storage object
jpayne@69:      * @return the "position"-th code unit.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual char16_t         setIndex(int32_t position) = 0;
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Sets the iterator to refer to the beginning of the code point
jpayne@69:      * that contains the "position"-th code unit
jpayne@69:      * in the text-storage object the iterator refers to, and
jpayne@69:      * returns that code point.
jpayne@69:      * The current position is adjusted to the beginning of the code point
jpayne@69:      * (its first code unit).
jpayne@69:      * @param position the "position"-th code unit in the text-storage object
jpayne@69:      * @return the "position"-th code point.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual UChar32       setIndex32(int32_t position) = 0;
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Returns the code unit the iterator currently refers to. 
jpayne@69:      * @return the current code unit. 
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual char16_t         current(void) const = 0;
jpayne@69:         
jpayne@69:     /**
jpayne@69:      * Returns the code point the iterator currently refers to.  
jpayne@69:      * @return the current code point.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual UChar32       current32(void) const = 0;
jpayne@69:         
jpayne@69:     /**
jpayne@69:      * Advances to the next code unit in the iteration range
jpayne@69:      * (toward endIndex()), and returns that code unit.  If there are
jpayne@69:      * no more code units to return, returns DONE.
jpayne@69:      * @return the next code unit.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual char16_t         next(void) = 0;
jpayne@69:         
jpayne@69:     /**
jpayne@69:      * Advances to the next code point in the iteration range
jpayne@69:      * (toward endIndex()), and returns that code point.  If there are
jpayne@69:      * no more code points to return, returns DONE.
jpayne@69:      * Note that iteration with "pre-increment" semantics is less
jpayne@69:      * efficient than iteration with "post-increment" semantics
jpayne@69:      * that is provided by next32PostInc().
jpayne@69:      * @return the next code point.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual UChar32       next32(void) = 0;
jpayne@69:         
jpayne@69:     /**
jpayne@69:      * Advances to the previous code unit in the iteration range
jpayne@69:      * (toward startIndex()), and returns that code unit.  If there are
jpayne@69:      * no more code units to return, returns DONE.  
jpayne@69:      * @return the previous code unit.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual char16_t         previous(void) = 0;
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Advances to the previous code point in the iteration range
jpayne@69:      * (toward startIndex()), and returns that code point.  If there are
jpayne@69:      * no more code points to return, returns DONE. 
jpayne@69:      * @return the previous code point. 
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual UChar32       previous32(void) = 0;
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Returns FALSE if there are no more code units or code points
jpayne@69:      * before the current position in the iteration range.
jpayne@69:      * This is used with previous() or previous32() in backward
jpayne@69:      * iteration.
jpayne@69:      * @return FALSE if there are no more code units or code points
jpayne@69:      * before the current position in the iteration range, return TRUE otherwise.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual UBool        hasPrevious() = 0;
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Returns the numeric index in the underlying text-storage
jpayne@69:      * object of the character returned by first().  Since it's
jpayne@69:      * possible to create an iterator that iterates across only
jpayne@69:      * part of a text-storage object, this number isn't
jpayne@69:      * necessarily 0.  
jpayne@69:      * @returns the numeric index in the underlying text-storage
jpayne@69:      * object of the character returned by first().
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     inline int32_t       startIndex(void) const;
jpayne@69:         
jpayne@69:     /**
jpayne@69:      * Returns the numeric index in the underlying text-storage
jpayne@69:      * object of the position immediately BEYOND the character
jpayne@69:      * returned by last().  
jpayne@69:      * @return the numeric index in the underlying text-storage
jpayne@69:      * object of the position immediately BEYOND the character
jpayne@69:      * returned by last().
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     inline int32_t       endIndex(void) const;
jpayne@69:         
jpayne@69:     /**
jpayne@69:      * Returns the numeric index in the underlying text-storage
jpayne@69:      * object of the character the iterator currently refers to
jpayne@69:      * (i.e., the character returned by current()).  
jpayne@69:      * @return the numeric index in the text-storage object of 
jpayne@69:      * the character the iterator currently refers to
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     inline int32_t       getIndex(void) const;
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Returns the length of the entire text in the underlying
jpayne@69:      * text-storage object.
jpayne@69:      * @return the length of the entire text in the text-storage object
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     inline int32_t           getLength() const;
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Moves the current position relative to the start or end of the
jpayne@69:      * iteration range, or relative to the current position itself.
jpayne@69:      * The movement is expressed in numbers of code units forward
jpayne@69:      * or backward by specifying a positive or negative delta.
jpayne@69:      * @param delta the position relative to origin. A positive delta means forward;
jpayne@69:      * a negative delta means backward.
jpayne@69:      * @param origin Origin enumeration {kStart, kCurrent, kEnd}
jpayne@69:      * @return the new position
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual int32_t      move(int32_t delta, EOrigin origin) = 0;
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Moves the current position relative to the start or end of the
jpayne@69:      * iteration range, or relative to the current position itself.
jpayne@69:      * The movement is expressed in numbers of code points forward
jpayne@69:      * or backward by specifying a positive or negative delta.
jpayne@69:      * @param delta the position relative to origin. A positive delta means forward;
jpayne@69:      * a negative delta means backward.
jpayne@69:      * @param origin Origin enumeration {kStart, kCurrent, kEnd}
jpayne@69:      * @return the new position
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69: #ifdef move32
jpayne@69:      // One of the system headers right now is sometimes defining a conflicting macro we don't use
jpayne@69: #undef move32
jpayne@69: #endif
jpayne@69:     virtual int32_t      move32(int32_t delta, EOrigin origin) = 0;
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Copies the text under iteration into the UnicodeString
jpayne@69:      * referred to by "result".  
jpayne@69:      * @param result Receives a copy of the text under iteration.  
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     virtual void            getText(UnicodeString&  result) = 0;
jpayne@69: 
jpayne@69: protected:
jpayne@69:     /**
jpayne@69:      * Empty constructor.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     CharacterIterator();
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Constructor, just setting the length field in this base class.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     CharacterIterator(int32_t length);
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Constructor, just setting the length and position fields in this base class.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     CharacterIterator(int32_t length, int32_t position);
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Constructor, just setting the length, start, end, and position fields in this base class.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     CharacterIterator(int32_t length, int32_t textBegin, int32_t textEnd, int32_t position);
jpayne@69:   
jpayne@69:     /**
jpayne@69:      * Copy constructor.
jpayne@69:      *
jpayne@69:      * @param that The CharacterIterator to be copied
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     CharacterIterator(const CharacterIterator &that);
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Assignment operator.  Sets this CharacterIterator to have the same behavior,
jpayne@69:      * as the one passed in.
jpayne@69:      * @param that The CharacterIterator passed in.
jpayne@69:      * @return the newly set CharacterIterator.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     CharacterIterator &operator=(const CharacterIterator &that);
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Base class text length field.
jpayne@69:      * Necessary this for correct getText() and hashCode().
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     int32_t textLength;
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Base class field for the current position.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     int32_t  pos;
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Base class field for the start of the iteration range.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     int32_t  begin;
jpayne@69: 
jpayne@69:     /**
jpayne@69:      * Base class field for the end of the iteration range.
jpayne@69:      * @stable ICU 2.0
jpayne@69:      */
jpayne@69:     int32_t  end;
jpayne@69: };
jpayne@69: 
jpayne@69: inline UBool
jpayne@69: ForwardCharacterIterator::operator!=(const ForwardCharacterIterator& that) const {
jpayne@69:     return !operator==(that);
jpayne@69: }
jpayne@69: 
jpayne@69: inline int32_t
jpayne@69: CharacterIterator::setToStart() {
jpayne@69:     return move(0, kStart);
jpayne@69: }
jpayne@69: 
jpayne@69: inline int32_t
jpayne@69: CharacterIterator::setToEnd() {
jpayne@69:     return move(0, kEnd);
jpayne@69: }
jpayne@69: 
jpayne@69: inline int32_t
jpayne@69: CharacterIterator::startIndex(void) const {
jpayne@69:     return begin;
jpayne@69: }
jpayne@69: 
jpayne@69: inline int32_t
jpayne@69: CharacterIterator::endIndex(void) const {
jpayne@69:     return end;
jpayne@69: }
jpayne@69: 
jpayne@69: inline int32_t
jpayne@69: CharacterIterator::getIndex(void) const {
jpayne@69:     return pos;
jpayne@69: }
jpayne@69: 
jpayne@69: inline int32_t
jpayne@69: CharacterIterator::getLength(void) const {
jpayne@69:     return textLength;
jpayne@69: }
jpayne@69: 
jpayne@69: U_NAMESPACE_END
jpayne@69: 
jpayne@69: #endif /* U_SHOW_CPLUSPLUS_API */
jpayne@69: 
jpayne@69: #endif