michael@0: /*
michael@0: **********************************************************************
michael@0: *   Copyright (C) 1998-2005, International Business Machines
michael@0: *   Corporation and others.  All Rights Reserved.
michael@0: **********************************************************************
michael@0: */
michael@0: 
michael@0: #ifndef UCHRITER_H
michael@0: #define UCHRITER_H
michael@0: 
michael@0: #include "unicode/utypes.h"
michael@0: #include "unicode/chariter.h"
michael@0: 
michael@0: /**
michael@0:  * \file 
michael@0:  * \brief C++ API: UChar Character Iterator
michael@0:  */
michael@0:  
michael@0: U_NAMESPACE_BEGIN
michael@0: 
michael@0: /**
michael@0:  * A concrete subclass of CharacterIterator that iterates over the
michael@0:  * characters (code units or code points) in a UChar array.
michael@0:  * It's possible not only to create an
michael@0:  * iterator that iterates over an entire UChar array, but also to
michael@0:  * create one that iterates over only a subrange of a UChar array
michael@0:  * (iterators over different subranges of the same UChar array don't
michael@0:  * compare equal).
michael@0:  * @see CharacterIterator
michael@0:  * @see ForwardCharacterIterator
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: class U_COMMON_API UCharCharacterIterator : public CharacterIterator {
michael@0: public:
michael@0:   /**
michael@0:    * Create an iterator over the UChar array referred to by "textPtr".
michael@0:    * The iteration range is 0 to <code>length-1</code>.
michael@0:    * text is only aliased, not adopted (the
michael@0:    * destructor will not delete it).
michael@0:    * @param textPtr The UChar array to be iterated over
michael@0:    * @param length The length of the UChar array
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   UCharCharacterIterator(const UChar* textPtr, int32_t length);
michael@0: 
michael@0:   /**
michael@0:    * Create an iterator over the UChar array referred to by "textPtr".
michael@0:    * The iteration range is 0 to <code>length-1</code>.
michael@0:    * text is only aliased, not adopted (the
michael@0:    * destructor will not delete it).
michael@0:    * The starting
michael@0:    * position is specified by "position". If "position" is outside the valid
michael@0:    * iteration range, the behavior of this object is undefined.
michael@0:    * @param textPtr The UChar array to be iteratd over
michael@0:    * @param length The length of the UChar array
michael@0:    * @param position The starting position of the iteration
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   UCharCharacterIterator(const UChar* textPtr, int32_t length,
michael@0:                          int32_t position);
michael@0: 
michael@0:   /**
michael@0:    * Create an iterator over the UChar array referred to by "textPtr".
michael@0:    * The iteration range is 0 to <code>end-1</code>.
michael@0:    * text is only aliased, not adopted (the
michael@0:    * destructor will not delete it).
michael@0:    * The starting
michael@0:    * position is specified by "position". If begin and end do not
michael@0:    * form a valid iteration range or "position" is outside the valid
michael@0:    * iteration range, the behavior of this object is undefined.
michael@0:    * @param textPtr The UChar array to be iterated over
michael@0:    * @param length The length of the UChar array
michael@0:    * @param textBegin  The begin position of the iteration range
michael@0:    * @param textEnd    The end position of the iteration range
michael@0:    * @param position    The starting position of the iteration
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   UCharCharacterIterator(const UChar* textPtr, int32_t length,
michael@0:                          int32_t textBegin,
michael@0:                          int32_t textEnd,
michael@0:                          int32_t position);
michael@0: 
michael@0:   /**
michael@0:    * Copy constructor.  The new iterator iterates over the same range
michael@0:    * of the same string as "that", and its initial position is the
michael@0:    * same as "that"'s current position.
michael@0:    * @param that The UCharCharacterIterator to be copied
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   UCharCharacterIterator(const UCharCharacterIterator&  that);
michael@0: 
michael@0:   /**
michael@0:    * Destructor.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual ~UCharCharacterIterator();
michael@0: 
michael@0:   /**
michael@0:    * Assignment operator.  *this is altered to iterate over the sane
michael@0:    * range of the same string as "that", and refers to the same
michael@0:    * character within that string as "that" does.
michael@0:    * @param that The object to be copied
michael@0:    * @return the newly created object
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   UCharCharacterIterator&
michael@0:   operator=(const UCharCharacterIterator&    that);
michael@0: 
michael@0:   /**
michael@0:    * Returns true if the iterators iterate over the same range of the
michael@0:    * same string and are pointing at the same character.
michael@0:    * @param that The ForwardCharacterIterator used to be compared for equality
michael@0:    * @return true if the iterators iterate over the same range of the
michael@0:    * same string and are pointing at the same character.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UBool          operator==(const ForwardCharacterIterator& that) const;
michael@0: 
michael@0:   /**
michael@0:    * Generates a hash code for this iterator.
michael@0:    * @return the hash code.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual int32_t         hashCode(void) const;
michael@0: 
michael@0:   /**
michael@0:    * Returns a new UCharCharacterIterator referring to the same
michael@0:    * character in the same range of the same string as this one.  The
michael@0:    * caller must delete the new iterator.
michael@0:    * @return the CharacterIterator newly created
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual CharacterIterator* clone(void) const;
michael@0: 
michael@0:   /**
michael@0:    * Sets the iterator to refer to the first code unit in its
michael@0:    * iteration range, and returns that code unit.
michael@0:    * This can be used to begin an iteration with next().
michael@0:    * @return the first code unit in its iteration range.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UChar         first(void);
michael@0: 
michael@0:   /**
michael@0:    * Sets the iterator to refer to the first code unit in its
michael@0:    * iteration range, returns that code unit, and moves the position
michael@0:    * to the second code unit. This is an alternative to setToStart()
michael@0:    * for forward iteration with nextPostInc().
michael@0:    * @return the first code unit in its iteration range
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UChar         firstPostInc(void);
michael@0: 
michael@0:   /**
michael@0:    * Sets the iterator to refer to the first code point in its
michael@0:    * iteration range, and returns that code unit,
michael@0:    * This can be used to begin an iteration with next32().
michael@0:    * Note that an iteration with next32PostInc(), beginning with,
michael@0:    * e.g., setToStart() or firstPostInc(), is more efficient.
michael@0:    * @return the first code point in its iteration range
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UChar32       first32(void);
michael@0: 
michael@0:   /**
michael@0:    * Sets the iterator to refer to the first code point in its
michael@0:    * iteration range, returns that code point, and moves the position
michael@0:    * to the second code point. This is an alternative to setToStart()
michael@0:    * for forward iteration with next32PostInc().
michael@0:    * @return the first code point in its iteration range.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UChar32       first32PostInc(void);
michael@0: 
michael@0:   /**
michael@0:    * Sets the iterator to refer to the last code unit in its
michael@0:    * iteration range, and returns that code unit.
michael@0:    * This can be used to begin an iteration with previous().
michael@0:    * @return the last code unit in its iteration range.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UChar         last(void);
michael@0: 
michael@0:   /**
michael@0:    * Sets the iterator to refer to the last code point in its
michael@0:    * iteration range, and returns that code unit.
michael@0:    * This can be used to begin an iteration with previous32().
michael@0:    * @return the last code point in its iteration range.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UChar32       last32(void);
michael@0: 
michael@0:   /**
michael@0:    * Sets the iterator to refer to the "position"-th code unit
michael@0:    * in the text-storage object the iterator refers to, and
michael@0:    * returns that code unit.
michael@0:    * @param position the position within the text-storage object
michael@0:    * @return the code unit
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UChar         setIndex(int32_t position);
michael@0: 
michael@0:   /**
michael@0:    * Sets the iterator to refer to the beginning of the code point
michael@0:    * that contains the "position"-th code unit
michael@0:    * in the text-storage object the iterator refers to, and
michael@0:    * returns that code point.
michael@0:    * The current position is adjusted to the beginning of the code point
michael@0:    * (its first code unit).
michael@0:    * @param position the position within the text-storage object
michael@0:    * @return the code unit
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UChar32       setIndex32(int32_t position);
michael@0: 
michael@0:   /**
michael@0:    * Returns the code unit the iterator currently refers to.
michael@0:    * @return the code unit the iterator currently refers to.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UChar         current(void) const;
michael@0: 
michael@0:   /**
michael@0:    * Returns the code point the iterator currently refers to.
michael@0:    * @return the code point the iterator currently refers to.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UChar32       current32(void) const;
michael@0: 
michael@0:   /**
michael@0:    * Advances to the next code unit in the iteration range (toward
michael@0:    * endIndex()), and returns that code unit.  If there are no more
michael@0:    * code units to return, returns DONE.
michael@0:    * @return the next code unit in the iteration range.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UChar         next(void);
michael@0: 
michael@0:   /**
michael@0:    * Gets the current code unit for returning and advances to the next code unit
michael@0:    * in the iteration range
michael@0:    * (toward endIndex()).  If there are
michael@0:    * no more code units to return, returns DONE.
michael@0:    * @return the current code unit.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UChar         nextPostInc(void);
michael@0: 
michael@0:   /**
michael@0:    * Advances to the next code point in the iteration range (toward
michael@0:    * endIndex()), and returns that code point.  If there are no more
michael@0:    * code points to return, returns DONE.
michael@0:    * Note that iteration with "pre-increment" semantics is less
michael@0:    * efficient than iteration with "post-increment" semantics
michael@0:    * that is provided by next32PostInc().
michael@0:    * @return the next code point in the iteration range.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UChar32       next32(void);
michael@0: 
michael@0:   /**
michael@0:    * Gets the current code point for returning and advances to the next code point
michael@0:    * in the iteration range
michael@0:    * (toward endIndex()).  If there are
michael@0:    * no more code points to return, returns DONE.
michael@0:    * @return the current point.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UChar32       next32PostInc(void);
michael@0: 
michael@0:   /**
michael@0:    * Returns FALSE if there are no more code units or code points
michael@0:    * at or after the current position in the iteration range.
michael@0:    * This is used with nextPostInc() or next32PostInc() in forward
michael@0:    * iteration.
michael@0:    * @return FALSE if there are no more code units or code points
michael@0:    * at or after the current position in the iteration range.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UBool        hasNext();
michael@0: 
michael@0:   /**
michael@0:    * Advances to the previous code unit in the iteration range (toward
michael@0:    * startIndex()), and returns that code unit.  If there are no more
michael@0:    * code units to return, returns DONE.
michael@0:    * @return the previous code unit in the iteration range.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UChar         previous(void);
michael@0: 
michael@0:   /**
michael@0:    * Advances to the previous code point in the iteration range (toward
michael@0:    * startIndex()), and returns that code point.  If there are no more
michael@0:    * code points to return, returns DONE.
michael@0:    * @return the previous code point in the iteration range.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UChar32       previous32(void);
michael@0: 
michael@0:   /**
michael@0:    * Returns FALSE if there are no more code units or code points
michael@0:    * before the current position in the iteration range.
michael@0:    * This is used with previous() or previous32() in backward
michael@0:    * iteration.
michael@0:    * @return FALSE if there are no more code units or code points
michael@0:    * before the current position in the iteration range.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UBool        hasPrevious();
michael@0: 
michael@0:   /**
michael@0:    * Moves the current position relative to the start or end of the
michael@0:    * iteration range, or relative to the current position itself.
michael@0:    * The movement is expressed in numbers of code units forward
michael@0:    * or backward by specifying a positive or negative delta.
michael@0:    * @param delta the position relative to origin. A positive delta means forward;
michael@0:    * a negative delta means backward.
michael@0:    * @param origin Origin enumeration {kStart, kCurrent, kEnd}
michael@0:    * @return the new position
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual int32_t      move(int32_t delta, EOrigin origin);
michael@0: 
michael@0:   /**
michael@0:    * Moves the current position relative to the start or end of the
michael@0:    * iteration range, or relative to the current position itself.
michael@0:    * The movement is expressed in numbers of code points forward
michael@0:    * or backward by specifying a positive or negative delta.
michael@0:    * @param delta the position relative to origin. A positive delta means forward;
michael@0:    * a negative delta means backward.
michael@0:    * @param origin Origin enumeration {kStart, kCurrent, kEnd}
michael@0:    * @return the new position
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual int32_t      move32(int32_t delta, EOrigin origin);
michael@0: 
michael@0:   /**
michael@0:    * Sets the iterator to iterate over a new range of text
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   void setText(const UChar* newText, int32_t newTextLength);
michael@0: 
michael@0:   /**
michael@0:    * Copies the UChar array under iteration into the UnicodeString
michael@0:    * referred to by "result".  Even if this iterator iterates across
michael@0:    * only a part of this string, the whole string is copied.
michael@0:    * @param result Receives a copy of the text under iteration.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual void            getText(UnicodeString& result);
michael@0: 
michael@0:   /**
michael@0:    * Return a class ID for this class (not really public)
michael@0:    * @return a class ID for this class
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   static UClassID         U_EXPORT2 getStaticClassID(void);
michael@0: 
michael@0:   /**
michael@0:    * Return a class ID for this object (not really public)
michael@0:    * @return a class ID for this object.
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   virtual UClassID        getDynamicClassID(void) const;
michael@0: 
michael@0: protected:
michael@0:   /**
michael@0:    * Protected constructor
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   UCharCharacterIterator();
michael@0:   /**
michael@0:    * Protected member text
michael@0:    * @stable ICU 2.0
michael@0:    */
michael@0:   const UChar*            text;
michael@0: 
michael@0: };
michael@0: 
michael@0: U_NAMESPACE_END
michael@0: #endif