collationelementiterator.java

java源代码 请看看啊 提点宝贵的意见

     * original string, not an offset into its corresponding sequence of
     * collation elements).  The value returned by the next call to next()
     * will be the collation element corresponding to the specified position
     * in the text.  If that position is in the middle of a contracting
     * character sequence, the result of the next call to next() is the
     * collation element for that sequence.  This means that getOffset()
     * is not guaranteed to return the same value as was passed to a preceding
     * call to setOffset().
     *
     * @param newOffset The new character offset into the original text.
     * @since 1.2
     */
    public void setOffset(int newOffset)
    {
        if (text != null) {
            if (newOffset < text.getBeginIndex()
                || newOffset >= text.getEndIndex()) {
                    text.setIndexOnly(newOffset);
            } else {
                char c = text.setIndex(newOffset);
                // if the desired character isn't used in a contracting character
                // sequence, bypass all the backing-up logic-- we're sitting on
                // the right character already
                if (ordering.usedInContractSeq(c)) {
                    // walk backwards through the string until we see a character
                    // that DOESN'T participate in a contracting character sequence
                    while (ordering.usedInContractSeq(c)) {
                        c = text.previous();
                    }
                    // now walk forward using this object's next() method until
                    // we pass the starting point and set our current position
                    // to the beginning of the last "character" before or at
                    // our starting position
                    int last = text.getIndex();
                    while (text.getIndex() <= newOffset) {
                        last = text.getIndex();
                        next();
                    }
                    text.setIndexOnly(last);
		    // we don't need this, since last is the last index 
		    // that is the starting of the contraction which encompass
		    // newOffset 
		    // text.previous();
                }
            }
        }
        buffer = null;
        expIndex = 0;
        swapOrder = 0;
    }

    /**
     * Returns the character offset in the original text corresponding to the next
     * collation element.  (That is, getOffset() returns the position in the text
     * corresponding to the collation element that will be returned by the next
     * call to next().)  This value will always be the index of the FIRST character
     * corresponding to the collation element (a contracting character sequence is
     * when two or more characters all correspond to the same collation element).
     * This means if you do setOffset(x) followed immediately by getOffset(), getOffset()
     * won't necessarily return x.
     *
     * @return The character offset in the original text corresponding to the collation
     * element that will be returned by the next call to next().
     * @since 1.2
     */
    public int getOffset()
    {
        return (text != null) ? text.getIndex() : 0;
    }


    /**
     * Return the maximum length of any expansion sequences that end
     * with the specified comparison order.
     * @param order a collation order returned by previous or next.
     * @return the maximum length of any expansion sequences ending
     *         with the specified order.
     * @since 1.2
     */
    public int getMaxExpansion(int order)
    {
        return ordering.getMaxExpansion(order);
    }

    /**
     * Set a new string over which to iterate.
     *
     * @param source  the new source text
     * @since 1.2
     */
    public void setText(String source)
    {
        buffer = null;
        swapOrder = 0;
        expIndex = 0;
        Normalizer.Mode mode =
            NormalizerUtilities.toNormalizerMode(owner.getDecomposition());
        if (text == null) {
            text = new Normalizer(source, mode);
        } else {
            text.setMode(mode);
            text.setText(source);
        }
    }

    /**
     * Set a new string over which to iterate.
     *
     * @param source  the new source text.
     * @since 1.2
     */
    public void setText(CharacterIterator source)
    {
        buffer = null;
        swapOrder = 0;
        expIndex = 0;
        Normalizer.Mode mode =
            NormalizerUtilities.toNormalizerMode(owner.getDecomposition());
        if (text == null) {
            text = new Normalizer(source, mode);
        } else {
            text.setMode(mode);
            text.setText(source);
        }
    }

    //============================================================
    // privates
    //============================================================

    /**
     * Determine if a character is a Thai vowel (which sorts after
     * its base consonant).
     */
    private final static boolean isThaiPreVowel(char ch) {
        return (ch >= '\u0e40') && (ch <= '\u0e44');
    }

    /**
     * Determine if a character is a Thai base consonant
     */
    private final static boolean isThaiBaseConsonant(char ch) {
        return (ch >= '\u0e01') && (ch <= '\u0e2e');
    }

    /**
     * Determine if a character is a Lao vowel (which sorts after
     * its base consonant).
     */
    private final static boolean isLaoPreVowel(char ch) {
        return (ch >= '\u0ec0') && (ch <= '\u0ec4');
    }

    /**
     * Determine if a character is a Lao base consonant
     */
    private final static boolean isLaoBaseConsonant(char ch) {
        return (ch >= '\u0e81') && (ch <= '\u0eae');
    }

    /**
     * This method produces a buffer which contains the collation
     * elements for the two characters, with colFirst's values preceding
     * another character's.  Presumably, the other character precedes colFirst
     * in logical order (otherwise you wouldn't need this method would you?).
     * The assumption is that the other char's value(s) have already been
     * computed.  If this char has a single element it is passed to this
     * method as lastValue, and lastExpansion is null.  If it has an
     * expansion it is passed in lastExpansion, and colLastValue is ignored.
     */
    private int[] makeReorderedBuffer(char colFirst,
                                      int lastValue,
                                      int[] lastExpansion,
                                      boolean forward) {

        int[] result;

        int firstValue = ordering.getUnicodeOrder(colFirst);
        if (firstValue >= RuleBasedCollator.CONTRACTCHARINDEX) {
            firstValue = forward? nextContractChar(colFirst) : prevContractChar(colFirst);
        }

        int[] firstExpansion = null;
        if (firstValue >= RuleBasedCollator.EXPANDCHARINDEX) {
            firstExpansion = ordering.getExpandValueList(firstValue);
        }

        if (!forward) {
            int temp1 = firstValue;
            firstValue = lastValue;
            lastValue = temp1;
            int[] temp2 = firstExpansion;
            firstExpansion = lastExpansion;
            lastExpansion = temp2;
        }

        if (firstExpansion == null && lastExpansion == null) {
            result = new int [2];
            result[0] = firstValue;
            result[1] = lastValue;
        }
        else {
            int firstLength = firstExpansion==null? 1 : firstExpansion.length;
            int lastLength = lastExpansion==null? 1 : lastExpansion.length;
            result = new int[firstLength + lastLength];

            if (firstExpansion == null) {
                result[0] = firstValue;
            }
            else {
                System.arraycopy(firstExpansion, 0, result, 0, firstLength);
            }

            if (lastExpansion == null) {
                result[firstLength] = lastValue;
            }
            else {
                System.arraycopy(lastExpansion, 0, result, firstLength, lastLength);
            }
        }

        return result;
    }

    /**
     *  Check if a comparison order is ignorable.
     *  @return true if a character is ignorable, false otherwise.
     */
    final static boolean isIgnorable(int order)
    {
        return ((primaryOrder(order) == 0) ? true : false);
    }

    /**
     * Get the ordering priority of the next contracting character in the
     * string.
     * @param ch the starting character of a contracting character token
     * @return the next contracting character's ordering.  Returns NULLORDER
     * if the end of string is reached.
     */
    private int nextContractChar(char ch)
    {
        // First get the ordering of this single character,
        // which is always the first element in the list
        Vector list = ordering.getContractValues(ch);
        EntryPair pair = (EntryPair)list.firstElement();
        int order = pair.value;

        // find out the length of the longest contracting character sequence in the list.
        // There's logic in the builder code to make sure the longest sequence is always
        // the last.
        pair = (EntryPair)list.lastElement();
        int maxLength = pair.entryName.length();

        // (the Normalizer is cloned here so that the seeking we do in the next loop
        // won't affect our real position in the text)
        Normalizer tempText = (Normalizer)text.clone();

        // extract the next maxLength characters in the string (we have to do this using the
        // Normalizer to ensure that our offsets correspond to those the rest of the
        // iterator is using) and store it in "fragment".
        tempText.previous();
        key.setLength(0);
        char c = tempText.next();
        while (maxLength > 0 && c != Normalizer.DONE) {
            key.append(c);
            --maxLength;
            c = tempText.next();
        }
        String fragment = key.toString();

        // now that we have that fragment, iterate through this list looking for the
        // longest sequence that matches the characters in the actual text.  (maxLength
        // is used here to keep track of the length of the longest sequence)
        // Upon exit from this loop, maxLength will contain the length of the matching
        // sequence and order will contain the collation-element value corresponding
        // to this sequence
        maxLength = 1;
        for (int i = list.size() - 1; i > 0; i--) {
            pair = (EntryPair)list.elementAt(i);
            if (!pair.fwd)
                continue;

            if (fragment.startsWith(pair.entryName) && pair.entryName.length()
                    > maxLength) {
                maxLength = pair.entryName.length();
                order = pair.value;
            }
        }

        // seek our current iteration position to the end of the matching sequence
        // and return the appropriate collation-element value (if there was no matching
        // sequence, we're already seeked to the right position and order already contains
        // the correct collation-element value for the single character)
        while (maxLength > 1) {
            text.next();
            --maxLength;
        }
        return order;
    }

    /**
     * Get the ordering priority of the previous contracting character in the
     * string.
     * @param ch the starting character of a contracting character token
     * @return the next contracting character's ordering.  Returns NULLORDER
     * if the end of string is reached.
     */
    private int prevContractChar(char ch)
    {
        // This function is identical to nextContractChar(), except that we've
        // switched things so that the next() and previous() calls on the Normalizer
        // are switched and so that we skip entry pairs with the fwd flag turned on
        // rather than off.  Notice that we still use append() and startsWith() when
        // working on the fragment.  This is because the entry pairs that are used
        // in reverse iteration have their names reversed already.
        Vector list = ordering.getContractValues(ch);
        EntryPair pair = (EntryPair)list.firstElement();
        int order = pair.value;

        pair = (EntryPair)list.lastElement();
        int maxLength = pair.entryName.length();

        Normalizer tempText = (Normalizer)text.clone();

        tempText.next();
        key.setLength(0);
        char c = tempText.previous();
        while (maxLength > 0 && c != Normalizer.DONE) {
            key.append(c);
            --maxLength;
            c = tempText.previous();
        }
        String fragment = key.toString();

        maxLength = 1;
        for (int i = list.size() - 1; i > 0; i--) {
            pair = (EntryPair)list.elementAt(i);
            if (pair.fwd)
                continue;

            if (fragment.startsWith(pair.entryName) && pair.entryName.length()
                    > maxLength) {
                maxLength = pair.entryName.length();
                order = pair.value;
            }
        }

        while (maxLength > 1) {
            text.previous();
            --maxLength;
        }
        return order;
    }

    final static int UNMAPPEDCHARVALUE = 0x7FFF0000;

    private Normalizer text = null;
    private int[] buffer = null;
    private int expIndex = 0;
    private StringBuffer key = new StringBuffer(5);
    private int swapOrder = 0;
    private RBCollationTables ordering;
    private RuleBasedCollator owner;
}