rulebasedcollator.java

gcc的组建

				   secondary_seq, tertiary_seq,
				   equality_seq, elt.expansionOrdering, elt.ignore));
      }

    this.inverseAccentComparison = inverseComparisons; 

    ce_table = v.toArray();

    last_primary_value = primary_seq+1;
    last_tertiary_value = last_tertiary_seq+1;
  }

  /**
   * Build a tree where all keys are the texts of collation elements and data is
   * the collation element itself. The tree is used when extracting all prefix
   * for a given text.
   */
  private void buildPrefixAccess()
  {
    prefix_tree = new HashMap();

    for (int i = 0; i < ce_table.length; i++)
      {
	CollationElement e = (CollationElement) ce_table[i];

	prefix_tree.put(e.key, e);
      }
  }

  /**
   * This method returns an integer which indicates whether the first
   * specified <code>String</code> is less than, greater than, or equal to
   * the second.  The value depends not only on the collation rules in
   * effect, but also the strength and decomposition settings of this object.
   *
   * @param source The first <code>String</code> to compare.
   * @param target A second <code>String</code> to compare to the first.
   *
   * @return A negative integer if source &lt; target, a positive integer
   * if source &gt; target, or 0 if source == target.
   */
  public int compare(String source, String target)
  {
    CollationElementIterator cs, ct;
    CollationElement ord1block = null;
    CollationElement ord2block = null;
    boolean advance_block_1 = true;
    boolean advance_block_2 = true;

    cs = getCollationElementIterator(source);
    ct = getCollationElementIterator(target);

    for(;;)
      {
	int ord1;
	int ord2;

	/*
	 * We have to check whether the characters are ignorable.
	 * If it is the case then forget them. 
	 */
	if (advance_block_1)
	  {
	    ord1block = cs.nextBlock();
	    if (ord1block != null && ord1block.ignore)
	      continue;
	  }
	
	if (advance_block_2)
	  {
	    ord2block = ct.nextBlock();
	    if (ord2block != null && ord2block.ignore)
	      {
	        advance_block_1 = false;
	        continue;
	      }
	 }
	else
	  advance_block_2 = true;

	if (!advance_block_1)
	  advance_block_1 = true;

	if (ord1block != null)
	  ord1 = ord1block.getValue();
	else
	  {
	    if (ord2block == null)
	      return 0;
	    return -1;
	  }

	if (ord2block == null)
	  return 1;
	
	ord2 = ord2block.getValue();
	
	// We know chars are totally equal, so skip
        if (ord1 == ord2)
	  {
	    if (getStrength() == IDENTICAL)
	      if (!ord1block.key.equals(ord2block.key))
		return ord1block.key.compareTo(ord2block.key);
	    continue;
	  }

        // Check for primary strength differences
        int prim1 = CollationElementIterator.primaryOrder(ord1); 
        int prim2 = CollationElementIterator.primaryOrder(ord2); 
	
	if (prim1 == 0 && getStrength() < TERTIARY)
	  {
            advance_block_2 = false;
	    continue;
	  }
	else if (prim2 == 0 && getStrength() < TERTIARY)
	  {
	    advance_block_1 = false;
	    continue;
	  }

        if (prim1 < prim2)
          return -1;
        else if (prim1 > prim2)
          return 1;
        else if (getStrength() == PRIMARY)
          continue;

        // Check for secondary strength differences
        int sec1 = CollationElementIterator.secondaryOrder(ord1);
        int sec2 = CollationElementIterator.secondaryOrder(ord2);

	if (sec1 < sec2)
          return -1;
        else if (sec1 > sec2)
          return 1;
        else if (getStrength() == SECONDARY)
          continue;

        // Check for tertiary differences
        int tert1 = CollationElementIterator.tertiaryOrder(ord1);
        int tert2 = CollationElementIterator.tertiaryOrder(ord2);

        if (tert1 < tert2)
          return -1;
        else if (tert1 > tert2)
          return 1;
	else if (getStrength() == TERTIARY)
	  continue;

	// Apparently JDK does this (at least for my test case).
	return ord1block.key.compareTo(ord2block.key);    
      }
  }

  /**
   * This method tests this object for equality against the specified 
   * object.  This will be true if and only if the specified object is
   * another reference to this object.
   *
   * @param obj The <code>Object</code> to compare against this object.
   *
   * @return <code>true</code> if the specified object is equal to this object,
   * <code>false</code> otherwise.
   */
  public boolean equals(Object obj)
  {
    if (obj == this)
      return true;
    else
      return false;
  }

  /**
   * This method builds a default collation element without invoking
   * the database created from the rules passed to the constructor.
   *
   * @param c Character which needs a collation element.
   * @return A valid brand new CollationElement instance.
   */
  CollationElement getDefaultElement(char c)
  {
    int v;

    // Preliminary support for generic accent sorting inversion (I don't know if all
    // characters in the range should be sorted backward). This is the place
    // to fix this if needed.
    if (inverseAccentComparison && (c >= 0x02B9 && c <= 0x0361))
      v = 0x0361 - ((int) c - 0x02B9);
    else
      v = (short) c;
    return new CollationElement("" + c, last_primary_value + v,
				(short) 0, (short) 0, (short) 0, null, false);
  }

  /**
   * This method builds a default collation element for an accented character
   * without invoking the database created from the rules passed to the constructor.
   *
   * @param c Character which needs a collation element.
   * @return A valid brand new CollationElement instance.
   */
  CollationElement getDefaultAccentedElement(char c)
  {
    int v;

    // Preliminary support for generic accent sorting inversion (I don't know if all
    // characters in the range should be sorted backward). This is the place
    // to fix this if needed.
    if (inverseAccentComparison && (c >= 0x02B9 && c <= 0x0361))
      v = 0x0361 - ((int) c - 0x02B9);
    else
      v = (short) c;
    return new CollationElement("" + c, (short) 0,
				(short) 0, (short) (last_tertiary_value + v), (short) 0, null, false);
  }

  /**
   * This method returns an instance for <code>CollationElementIterator</code>
   * for the specified <code>String</code> under the collation rules for this
   * object.
   *
   * @param source The <code>String</code> to return the
   * <code>CollationElementIterator</code> instance for.
   *
   * @return A <code>CollationElementIterator</code> for the specified
   * <code>String</code>.
   */
  public CollationElementIterator getCollationElementIterator(String source)
  {
    return new CollationElementIterator(this, source);
  }

  /**
   * This method returns an instance of <code>CollationElementIterator</code>
   * for the <code>String</code> represented by the specified
   * <code>CharacterIterator</code>.
   *
   * @param source The <code>CharacterIterator</code> with the desired <code>String</code>.
   *
   * @return A <code>CollationElementIterator</code> for the specified <code>String</code>.
   */
  public CollationElementIterator getCollationElementIterator(CharacterIterator source)
  {
    StringBuffer expand = new StringBuffer("");
    
    // Right now we assume that we will read from the beginning of the string.
    for (char c = source.first();
	 c != CharacterIterator.DONE;
	 c = source.next())
      decomposeCharacter(c, expand);

    return getCollationElementIterator(expand.toString());
  }

  /**
   * This method returns an instance of <code>CollationKey</code> for the
   * specified <code>String</code>.  The object returned will have a
   * more efficient mechanism for its comparison function that could
   * provide speed benefits if multiple comparisons are performed, such
   * as during a sort.
   *
   * @param source The <code>String</code> to create a <code>CollationKey</code> for.
   *
   * @return A <code>CollationKey</code> for the specified <code>String</code>.
   */
  public CollationKey getCollationKey(String source)
  {
    CollationElementIterator cei = getCollationElementIterator(source);
    ArrayList vect = new ArrayList();

    int ord = cei.next();
    cei.reset(); //set to start of string

    while (ord != CollationElementIterator.NULLORDER)
      {
	// If the primary order is null, it means this is an ignorable
	// character.
	if (CollationElementIterator.primaryOrder(ord) == 0)
	  {
            ord = cei.next();
	    continue;
	  }
        switch (getStrength())
          {
            case PRIMARY:
	      ord = CollationElementIterator.primaryOrder(ord);
	      break;
	      
            case SECONDARY:
	      ord = CollationElementIterator.primaryOrder(ord) << 8;
	      ord |= CollationElementIterator.secondaryOrder(ord);

            default:
               break;
          }

        vect.add(new Integer(ord)); 
	ord = cei.next(); //increment to next key
      }

    Object[] objarr = vect.toArray();
    byte[] key = new byte[objarr.length * 4];

    for (int i = 0; i < objarr.length; i++)
      {
        int j = ((Integer) objarr[i]).intValue();
        key [i * 4] = (byte) ((j & 0xFF000000) >> 24);
        key [i * 4 + 1] = (byte) ((j & 0x00FF0000) >> 16);
        key [i * 4 + 2] = (byte) ((j & 0x0000FF00) >> 8);
        key [i * 4 + 3] = (byte) (j & 0x000000FF);
      }

    return new CollationKey(this, source, key);
  }

  /**
   * This method returns a <code>String</code> containing the collation rules
   * for this object.
   *
   * @return The collation rules for this object.
   */
  public String getRules()
  {
    return rules;
  }

  /**
   * This method returns a hash value for this object.
   *
   * @return A hash value for this object.
   */
  public int hashCode()
  {
    return System.identityHashCode(this);
  }
}