Use Tree Navigation
public class

UnicodeSet

extends Object
implements UnicodeMatcher
/*
 * Copyright (c) 2005, 2006, Oracle and/or its affiliates. All rights reserved.
 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
 *
 * This code is free software; you can redistribute it and/or modify it
 * under the terms of the GNU General Public License version 2 only, as
 * published by the Free Software Foundation.  Oracle designates this
 * particular file as subject to the "Classpath" exception as provided
 * by Oracle in the LICENSE file that accompanied this code.
 *
 * This code is distributed in the hope that it will be useful, but WITHOUT
 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
 * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
 * version 2 for more details (a copy is included in the LICENSE file that
 * accompanied this code).
 *
 * You should have received a copy of the GNU General Public License version
 * 2 along with this work; if not, write to the Free Software Foundation,
 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
 *
 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
 * or visit www.oracle.com if you need additional information or have any
 * questions.
 */


/*
 *******************************************************************************
 * (C) Copyright IBM Corp. 1996-2005 - All Rights Reserved                     *
 *                                                                             *
 * The original version of this source code and documentation is copyrighted   *
 * and owned by IBM, These materials are provided under terms of a License     *
 * Agreement between IBM and Sun. This technology is protected by multiple     *
 * US and International patents. This notice and attribution to IBM may not    *
 * to removed.                                                                 *
 *******************************************************************************
 */


package sun.text.normalizer;

import java.text.ParsePosition;
import java.util.Map;
import java.util.HashMap;
import java.util.TreeSet;
import java.util.Iterator;
import java.util.Collection;

/**
 * A mutable set of Unicode characters and multicharacter strings.  Objects of this class
 * represent <em>character classes</em> used in regular expressions.
 * A character specifies a subset of Unicode code points.  Legal
 * code points are U+0000 to U+10FFFF, inclusive.
 *
 * <p>The UnicodeSet class is not designed to be subclassed.
 *
 * <p><code>UnicodeSet</code> supports two APIs. The first is the
 * <em>operand</em> API that allows the caller to modify the value of
 * a <code>UnicodeSet</code> object. It conforms to Java 2's
 * <code>java.util.Set</code> interface, although
 * <code>UnicodeSet</code> does not actually implement that
 * interface. All methods of <code>Set</code> are supported, with the
 * modification that they take a character range or single character
 * instead of an <code>Object</code>, and they take a
 * <code>UnicodeSet</code> instead of a <code>Collection</code>.  The
 * operand API may be thought of in terms of boolean logic: a boolean
 * OR is implemented by <code>add</code>, a boolean AND is implemented
 * by <code>retain</code>, a boolean XOR is implemented by
 * <code>complement</code> taking an argument, and a boolean NOT is
 * implemented by <code>complement</code> with no argument.  In terms
 * of traditional set theory function names, <code>add</code> is a
 * union, <code>retain</code> is an intersection, <code>remove</code>
 * is an asymmetric difference, and <code>complement</code> with no
 * argument is a set complement with respect to the superset range
 * <code>MIN_VALUE-MAX_VALUE</code>
 *
 * <p>The second API is the
 * <code>applyPattern()</code>/<code>toPattern()</code> API from the
 * <code>java.text.Format</code>-derived classes.  Unlike the
 * methods that add characters, add categories, and control the logic
 * of the set, the method <code>applyPattern()</code> sets all
 * attributes of a <code>UnicodeSet</code> at once, based on a
 * string pattern.
 *
 * <p><b>Pattern syntax</b></p>
 *
 * Patterns are accepted by the constructors and the
 * <code>applyPattern()</code> methods and returned by the
 * <code>toPattern()</code> method.  These patterns follow a syntax
 * similar to that employed by version 8 regular expression character
 * classes.  Here are some simple examples:
 *
 * <blockquote>
 *   <table>
 *     <tr align="top">
 *       <td nowrap valign="top" align="left"><code>[]</code></td>
 *       <td valign="top">No characters</td>
 *     </tr><tr align="top">
 *       <td nowrap valign="top" align="left"><code>[a]</code></td>
 *       <td valign="top">The character 'a'</td>
 *     </tr><tr align="top">
 *       <td nowrap valign="top" align="left"><code>[ae]</code></td>
 *       <td valign="top">The characters 'a' and 'e'</td>
 *     </tr>
 *     <tr>
 *       <td nowrap valign="top" align="left"><code>[a-e]</code></td>
 *       <td valign="top">The characters 'a' through 'e' inclusive, in Unicode code
 *       point order</td>
 *     </tr>
 *     <tr>
 *       <td nowrap valign="top" align="left"><code>[\\u4E01]</code></td>
 *       <td valign="top">The character U+4E01</td>
 *     </tr>
 *     <tr>
 *       <td nowrap valign="top" align="left"><code>[a{ab}{ac}]</code></td>
 *       <td valign="top">The character 'a' and the multicharacter strings &quot;ab&quot; and
 *       &quot;ac&quot;</td>
 *     </tr>
 *     <tr>
 *       <td nowrap valign="top" align="left"><code>[\p{Lu}]</code></td>
 *       <td valign="top">All characters in the general category Uppercase Letter</td>
 *     </tr>
 *   </table>
 * </blockquote>
 *
 * Any character may be preceded by a backslash in order to remove any special
 * meaning.  White space characters, as defined by UCharacterProperty.isRuleWhiteSpace(), are
 * ignored, unless they are escaped.
 *
 * <p>Property patterns specify a set of characters having a certain
 * property as defined by the Unicode standard.  Both the POSIX-like
 * "[:Lu:]" and the Perl-like syntax "\p{Lu}" are recognized.  For a
 * complete list of supported property patterns, see the User's Guide
 * for UnicodeSet at
 * <a href="http://oss.software.ibm.com/icu/userguide/unicodeSet.html">
 * http://oss.software.ibm.com/icu/userguide/unicodeSet.html</a>.
 * Actual determination of property data is defined by the underlying
 * Unicode database as implemented by UCharacter.
 *
 * <p>Patterns specify individual characters, ranges of characters, and
 * Unicode property sets.  When elements are concatenated, they
 * specify their union.  To complement a set, place a '^' immediately
 * after the opening '['.  Property patterns are inverted by modifying
 * their delimiters; "[:^foo]" and "\P{foo}".  In any other location,
 * '^' has no special meaning.
 *
 * <p>Ranges are indicated by placing two a '-' between two
 * characters, as in "a-z".  This specifies the range of all
 * characters from the left to the right, in Unicode order.  If the
 * left character is greater than or equal to the
 * right character it is a syntax error.  If a '-' occurs as the first
 * character after the opening '[' or '[^', or if it occurs as the
 * last character before the closing ']', then it is taken as a
 * literal.  Thus "[a\\-b]", "[-ab]", and "[ab-]" all indicate the same
 * set of three characters, 'a', 'b', and '-'.
 *
 * <p>Sets may be intersected using the '&' operator or the asymmetric
 * set difference may be taken using the '-' operator, for example,
 * "[[:L:]&[\\u0000-\\u0FFF]]" indicates the set of all Unicode letters
 * with values less than 4096.  Operators ('&' and '|') have equal
 * precedence and bind left-to-right.  Thus
 * "[[:L:]-[a-z]-[\\u0100-\\u01FF]]" is equivalent to
 * "[[[:L:]-[a-z]]-[\\u0100-\\u01FF]]".  This only really matters for
 * difference; intersection is commutative.
 *
 * <table>
 * <tr valign=top><td nowrap><code>[a]</code><td>The set containing 'a'
 * <tr valign=top><td nowrap><code>[a-z]</code><td>The set containing 'a'
 * through 'z' and all letters in between, in Unicode order
 * <tr valign=top><td nowrap><code>[^a-z]</code><td>The set containing
 * all characters but 'a' through 'z',
 * that is, U+0000 through 'a'-1 and 'z'+1 through U+10FFFF
 * <tr valign=top><td nowrap><code>[[<em>pat1</em>][<em>pat2</em>]]</code>
 * <td>The union of sets specified by <em>pat1</em> and <em>pat2</em>
 * <tr valign=top><td nowrap><code>[[<em>pat1</em>]&[<em>pat2</em>]]</code>
 * <td>The intersection of sets specified by <em>pat1</em> and <em>pat2</em>
 * <tr valign=top><td nowrap><code>[[<em>pat1</em>]-[<em>pat2</em>]]</code>
 * <td>The asymmetric difference of sets specified by <em>pat1</em> and
 * <em>pat2</em>
 * <tr valign=top><td nowrap><code>[:Lu:] or \p{Lu}</code>
 * <td>The set of characters having the specified
 * Unicode property; in
 * this case, Unicode uppercase letters
 * <tr valign=top><td nowrap><code>[:^Lu:] or \P{Lu}</code>
 * <td>The set of characters <em>not</em> having the given
 * Unicode property
 * </table>
 *
 * <p><b>Warning</b>: you cannot add an empty string ("") to a UnicodeSet.</p>
 *
 * <p><b>Formal syntax</b></p>
 *
 * <blockquote>
 *   <table>
 *     <tr align="top">
 *       <td nowrap valign="top" align="right"><code>pattern :=  </code></td>
 *       <td valign="top"><code>('[' '^'? item* ']') |
 *       property</code></td>
 *     </tr>
 *     <tr align="top">
 *       <td nowrap valign="top" align="right"><code>item :=  </code></td>
 *       <td valign="top"><code>char | (char '-' char) | pattern-expr<br>
 *       </code></td>
 *     </tr>
 *     <tr align="top">
 *       <td nowrap valign="top" align="right"><code>pattern-expr :=  </code></td>
 *       <td valign="top"><code>pattern | pattern-expr pattern |
 *       pattern-expr op pattern<br>
 *       </code></td>
 *     </tr>
 *     <tr align="top">
 *       <td nowrap valign="top" align="right"><code>op :=  </code></td>
 *       <td valign="top"><code>'&amp;' | '-'<br>
 *       </code></td>
 *     </tr>
 *     <tr align="top">
 *       <td nowrap valign="top" align="right"><code>special :=  </code></td>
 *       <td valign="top"><code>'[' | ']' | '-'<br>
 *       </code></td>
 *     </tr>
 *     <tr align="top">
 *       <td nowrap valign="top" align="right"><code>char :=  </code></td>
 *       <td valign="top"><em>any character that is not</em><code> special<br>
 *       | ('\\' </code><em>any character</em><code>)<br>
 *       | ('&#92;u' hex hex hex hex)<br>
 *       </code></td>
 *     </tr>
 *     <tr align="top">
 *       <td nowrap valign="top" align="right"><code>hex :=  </code></td>
 *       <td valign="top"><em>any character for which
 *       </em><code>Character.digit(c, 16)</code><em>
 *       returns a non-negative result</em></td>
 *     </tr>
 *     <tr>
 *       <td nowrap valign="top" align="right"><code>property :=  </code></td>
 *       <td valign="top"><em>a Unicode property set pattern</td>
 *     </tr>
 *   </table>
 *   <br>
 *   <table border="1">
 *     <tr>
 *       <td>Legend: <table>
 *         <tr>
 *           <td nowrap valign="top"><code>a := b</code></td>
 *           <td width="20" valign="top">  </td>
 *           <td valign="top"><code>a</code> may be replaced by <code>b</code> </td>
 *         </tr>
 *         <tr>
 *           <td nowrap valign="top"><code>a?</code></td>
 *           <td valign="top"></td>
 *           <td valign="top">zero or one instance of <code>a</code><br>
 *           </td>
 *         </tr>
 *         <tr>
 *           <td nowrap valign="top"><code>a*</code></td>
 *           <td valign="top"></td>
 *           <td valign="top">one or more instances of <code>a</code><br>
 *           </td>
 *         </tr>
 *         <tr>
 *           <td nowrap valign="top"><code>a | b</code></td>
 *           <td valign="top"></td>
 *           <td valign="top">either <code>a</code> or <code>b</code><br>
 *           </td>
 *         </tr>
 *         <tr>
 *           <td nowrap valign="top"><code>'a'</code></td>
 *           <td valign="top"></td>
 *           <td valign="top">the literal string between the quotes </td>
 *         </tr>
 *       </table>
 *       </td>
 *     </tr>
 *   </table>
 * </blockquote>
 *
 * @author Alan Liu
 * @stable ICU 2.0
 */

public class UnicodeSet implements UnicodeMatcher {

   
private static final int LOW = 0x000000; // LOW <= all valid values. ZERO for codepoints
   
private static final int HIGH = 0x110000; // HIGH > all valid values. 10000 for code units.
                                             
// 110000 for codepoints

   
/**
     * Minimum value that can be stored in a UnicodeSet.
     * @stable ICU 2.0
     */

   
public static final int MIN_VALUE = LOW;

   
/**
     * Maximum value that can be stored in a UnicodeSet.
     * @stable ICU 2.0
     */

   
public static final int MAX_VALUE = HIGH - 1;

   
private int len;      // length used; list may be longer to minimize reallocs
   
private int[] list;   // MUST be terminated with HIGH
   
private int[] rangeList; // internal buffer
   
private int[] buffer; // internal buffer

   
// NOTE: normally the field should be of type SortedSet; but that is missing a public clone!!
   
// is not private so that UnicodeSetIterator can get access
   
TreeSet strings = new TreeSet();

   
/**
     * The pattern representation of this set.  This may not be the
     * most economical pattern.  It is the pattern supplied to
     * applyPattern(), with variables substituted and whitespace
     * removed.  For sets constructed without applyPattern(), or
     * modified using the non-pattern API, this string will be null,
     * indicating that toPattern() must generate a pattern
     * representation from the inversion list.
     */

   
private String pat = null;

   
private static final int START_EXTRA = 16;         // initial storage. Must be >= 0
   
private static final int GROW_EXTRA = START_EXTRA; // extra amount for growth. Must be >= 0

   
/**
     * A set of all characters _except_ the second through last characters of
     * certain ranges.  These ranges are ranges of characters whose
     * properties are all exactly alike, e.g. CJK Ideographs from
     * U+4E00 to U+9FA5.
     */

   
private static UnicodeSet INCLUSIONS = null;

   
//----------------------------------------------------------------
   
// Public API
   
//----------------------------------------------------------------

   
/**
     * Constructs an empty set.
     * @stable ICU 2.0
     */

   
public UnicodeSet() {
        list
= new int[1 + START_EXTRA];
        list
[len++] = HIGH;
   
}

   
/**
     * Constructs a set containing the given range. If <code>end >
     * start</code> then an empty set is created.
     *
     * @param start first character, inclusive, of range
     * @param end last character, inclusive, of range
     * @stable ICU 2.0
     */

   
public UnicodeSet(int start, int end) {
       
this();
        complement
(start, end);
   
}

   
/**
     * Constructs a set from the given pattern.  See the class description
     * for the syntax of the pattern language.  Whitespace is ignored.
     * @param pattern a string specifying what characters are in the set
     * @exception java.lang.IllegalArgumentException if the pattern contains
     * a syntax error.
     * @stable ICU 2.0
     */

   
public UnicodeSet(String pattern) {
       
this();
        applyPattern
(pattern, null, null, IGNORE_SPACE);
   
}

   
/**
     * Make this object represent the same set as <code>other</code>.
     * @param other a <code>UnicodeSet</code> whose value will be
     * copied to this object
     * @stable ICU 2.0
     */

   
public UnicodeSet set(UnicodeSet other) {
        list
= (int[]) other.list.clone();
        len
= other.len;
        pat
= other.pat;
        strings
= (TreeSet)other.strings.clone();
       
return this;
   
}

   
/**
     * Modifies this set to represent the set specified by the given pattern.
     * See the class description for the syntax of the pattern language.
     * Whitespace is ignored.
     * @param pattern a string specifying what characters are in the set
     * @exception java.lang.IllegalArgumentException if the pattern
     * contains a syntax error.
     * @stable ICU 2.0
     */

   
public final UnicodeSet applyPattern(String pattern) {
       
return applyPattern(pattern, null, null, IGNORE_SPACE);
   
}

   
/**
     * Append the <code>toPattern()</code> representation of a
     * string to the given <code>StringBuffer</code>.
     */

   
private static void _appendToPat(StringBuffer buf, String s, boolean escapeUnprintable) {
       
for (int i = 0; i < s.length(); i += UTF16.getCharCount(i)) {
            _appendToPat
(buf, UTF16.charAt(s, i), escapeUnprintable);
       
}
   
}

   
/**
     * Append the <code>toPattern()</code> representation of a
     * character to the given <code>StringBuffer</code>.
     */

   
private static void _appendToPat(StringBuffer buf, int c, boolean escapeUnprintable) {
       
if (escapeUnprintable && Utility.isUnprintable(c)) {
           
// Use hex escape notation (<backslash>uxxxx or <backslash>Uxxxxxxxx) for anything
           
// unprintable
           
if (Utility.escapeUnprintable(buf, c)) {
               
return;
           
}
       
}
       
// Okay to let ':' pass through
       
switch (c) {
       
case '[': // SET_OPEN:
       
case ']': // SET_CLOSE:
       
case '-': // HYPHEN:
       
case '^': // COMPLEMENT:
       
case '&': // INTERSECTION:
       
case '\\': //BACKSLASH:
       
case '{':
       
case '}':
       
case '$':
       
case ':':
            buf
.append('\\');
           
break;
       
default:
           
// Escape whitespace
           
if (UCharacterProperty.isRuleWhiteSpace(c)) {
                buf
.append('\\');
           
}
           
break;
       
}
        UTF16
.append(buf, c);
   
}

   
/**
     * Append a string representation of this set to result.  This will be
     * a cleaned version of the string passed to applyPattern(), if there
     * is one.  Otherwise it will be generated.
     */

   
private StringBuffer _toPattern(StringBuffer result,
                                   
boolean escapeUnprintable) {
       
if (pat != null) {
           
int i;
           
int backslashCount = 0;
           
for (i=0; i<pat.length(); ) {
               
int c = UTF16.charAt(pat, i);
                i
+= UTF16.getCharCount(c);
               
if (escapeUnprintable && Utility.isUnprintable(c)) {
                   
// If the unprintable character is preceded by an odd
                   
// number of backslashes, then it has been escaped.
                   
// Before unescaping it, we delete the final
                   
// backslash.
                   
if ((backslashCount % 2) == 1) {
                        result
.setLength(result.length() - 1);
                   
}
                   
Utility.escapeUnprintable(result, c);
                    backslashCount
= 0;
               
} else {
                    UTF16
.append(result, c);
                   
if (c == '\\') {
                       
++backslashCount;
                   
} else {
                        backslashCount
= 0;
                   
}
               
}
           
}
           
return result;
       
}

       
return _generatePattern(result, escapeUnprintable);
   
}

   
/**
     * Generate and append a string representation of this set to result.
     * This does not use this.pat, the cleaned up copy of the string
     * passed to applyPattern().
     * @stable ICU 2.0
     */

   
public StringBuffer _generatePattern(StringBuffer result,
                                         
boolean escapeUnprintable) {
        result
.append('[');

       
int count = getRangeCount();

       
// If the set contains at least 2 intervals and includes both
       
// MIN_VALUE and MAX_VALUE, then the inverse representation will
       
// be more economical.
       
if (count > 1 &&
            getRangeStart
(0) == MIN_VALUE &&
            getRangeEnd
(count-1) == MAX_VALUE) {

           
// Emit the inverse
            result
.append('^');

           
for (int i = 1; i < count; ++i) {
               
int start = getRangeEnd(i-1)+1;
               
int end = getRangeStart(i)-1;
                _appendToPat
(result, start, escapeUnprintable);
               
if (start != end) {
                   
if ((start+1) != end) {
                        result
.append('-');
                   
}
                    _appendToPat
(result, end, escapeUnprintable);
               
}
           
}
       
}

       
// Default; emit the ranges as pairs
       
else {
           
for (int i = 0; i < count; ++i) {
               
int start = getRangeStart(i);
               
int end = getRangeEnd(i);
                _appendToPat
(result, start, escapeUnprintable);
               
if (start != end) {
                   
if ((start+1) != end) {
                        result
.append('-');
                   
}
                    _appendToPat
(result, end, escapeUnprintable);
               
}
           
}
       
}

       
if (strings.size() > 0) {
           
Iterator it = strings.iterator();
           
while (it.hasNext()) {
                result
.append('{');
                _appendToPat
(result, (String) it.next(), escapeUnprintable);
                result
.append('}');
           
}
       
}
       
return result.append(']');
   
}

   
/**
     * Adds the specified range to this set if it is not already
     * present.  If this set already contains the specified range,
     * the call leaves this set unchanged.  If <code>end > start</code>
     * then an empty range is added, leaving the set unchanged.
     *
     * @param start first character, inclusive, of range to be added
     * to this set.
     * @param end last character, inclusive, of range to be added
     * to this set.
     * @stable ICU 2.0
     */

   
public UnicodeSet add(int start, int end) {
       
if (start < MIN_VALUE || start > MAX_VALUE) {
           
throw new IllegalArgumentException("Invalid code point U+" + Utility.hex(start, 6));
       
}
       
if (end < MIN_VALUE || end > MAX_VALUE) {
           
throw new IllegalArgumentException("Invalid code point U+" + Utility.hex(end, 6));
       
}
       
if (start < end) {
            add
(range(start, end), 2, 0);
       
} else if (start == end) {
            add
(start);
       
}
       
return this;
   
}

   
/**
     * Adds the specified character to this set if it is not already
     * present.  If this set already contains the specified character,
     * the call leaves this set unchanged.
     * @stable ICU 2.0
     */

   
public final UnicodeSet add(int c) {
       
if (c < MIN_VALUE || c > MAX_VALUE) {
           
throw new IllegalArgumentException("Invalid code point U+" + Utility.hex(c, 6));
       
}

       
// find smallest i such that c < list[i]
       
// if odd, then it is IN the set
       
// if even, then it is OUT of the set
       
int i = findCodePoint(c);

       
// already in set?
       
if ((i & 1) != 0) return this;

       
// HIGH is 0x110000
       
// assert(list[len-1] == HIGH);

       
// empty = [HIGH]
       
// [start_0, limit_0, start_1, limit_1, HIGH]

       
// [..., start_k-1, limit_k-1, start_k, limit_k, ..., HIGH]
       
//                             ^
       
//                             list[i]

       
// i == 0 means c is before the first range

       
if (c == list[i]-1) {
           
// c is before start of next range
            list
[i] = c;
           
// if we touched the HIGH mark, then add a new one
           
if (c == MAX_VALUE) {
                ensureCapacity
(len+1);
                list
[len++] = HIGH;
           
}
           
if (i > 0 && c == list[i-1]) {
               
// collapse adjacent ranges

               
// [..., start_k-1, c, c, limit_k, ..., HIGH]
               
//                     ^
               
//                     list[i]
               
System.arraycopy(list, i+1, list, i-1, len-i-1);
                len
-= 2;
           
}
       
}

       
else if (i > 0 && c == list[i-1]) {
           
// c is after end of prior range
            list
[i-1]++;
           
// no need to chcek for collapse here
       
}

       
else {
           
// At this point we know the new char is not adjacent to
           
// any existing ranges, and it is not 10FFFF.


           
// [..., start_k-1, limit_k-1, start_k, limit_k, ..., HIGH]
           
//                             ^
           
//                             list[i]

           
// [..., start_k-1, limit_k-1, c, c+1, start_k, limit_k, ..., HIGH]
           
//                             ^
           
//                             list[i]

           
// Don't use ensureCapacity() to save on copying.
           
// NOTE: This has no measurable impact on performance,
           
// but it might help in some usage patterns.
           
if (len+2 > list.length) {
               
int[] temp = new int[len + 2 + GROW_EXTRA];
               
if (i != 0) System.arraycopy(list, 0, temp, 0, i);
               
System.arraycopy(list, i, temp, i+2, len-i);
                list
= temp;
           
} else {
               
System.arraycopy(list, i, list, i+2, len-i);
           
}

            list
[i] = c;
            list
[i+1] = c+1;
            len
+= 2;
       
}

        pat
= null;
       
return this;
   
}

   
/**
     * Adds the specified multicharacter to this set if it is not already
     * present.  If this set already contains the multicharacter,
     * the call leaves this set unchanged.
     * Thus "ch" => {"ch"}
     * <br><b>Warning: you cannot add an empty string ("") to a UnicodeSet.</b>
     * @param s the source string
     * @return this object, for chaining
     * @stable ICU 2.0
     */

   
public final UnicodeSet add(String s) {

       
int cp = getSingleCP(s);
       
if (cp < 0) {
            strings
.add(s);
            pat
= null;
       
} else {
            add
(cp, cp);
       
}
       
return this;
   
}

   
/**
     * @return a code point IF the string consists of a single one.
     * otherwise returns -1.
     * @param string to test
     */

   
private static int getSingleCP(String s) {
       
if (s.length() < 1) {
           
throw new IllegalArgumentException("Can't use zero-length strings in UnicodeSet");
       
}
       
if (s.length() > 2) return -1;
       
if (s.length() == 1) return s.charAt(0);

       
// at this point, len = 2
       
int cp = UTF16.charAt(s, 0);
       
if (cp > 0xFFFF) { // is surrogate pair
           
return cp;
       
}
       
return -1;
   
}

   
/**
     * Complements the specified range in this set.  Any character in
     * the range will be removed if it is in this set, or will be
     * added if it is not in this set.  If <code>end > start</code>
     * then an empty range is complemented, leaving the set unchanged.
     *
     * @param start first character, inclusive, of range to be removed
     * from this set.
     * @param end last character, inclusive, of range to be removed
     * from this set.
     * @stable ICU 2.0
     */

   
public UnicodeSet complement(int start, int end) {
       
if (start < MIN_VALUE || start > MAX_VALUE) {
           
throw new IllegalArgumentException("Invalid code point U+" + Utility.hex(start, 6));
       
}
       
if (end < MIN_VALUE || end > MAX_VALUE) {
           
throw new IllegalArgumentException("Invalid code point U+" + Utility.hex(end, 6));
       
}
       
if (start <= end) {
            xor
(range(start, end), 2, 0);
       
}
        pat
= null;
       
return this;
   
}

   
/**
     * This is equivalent to
     * <code>complement(MIN_VALUE, MAX_VALUE)</code>.
     * @stable ICU 2.0
     */

   
public UnicodeSet complement() {
       
if (list[0] == LOW) {
           
System.arraycopy(list, 1, list, 0, len-1);
           
--len;
       
} else {
            ensureCapacity
(len+1);
           
System.arraycopy(list, 0, list, 1, len);
            list
[0] = LOW;
           
++len;
       
}
        pat
= null;
       
return this;
   
}

   
/**
     * Returns true if this set contains the given character.
     * @param c character to be checked for containment
     * @return true if the test condition is met
     * @stable ICU 2.0
     */

   
public boolean contains(int c) {
       
if (c < MIN_VALUE || c > MAX_VALUE) {
           
throw new IllegalArgumentException("Invalid code point U+" + Utility.hex(c, 6));
       
}

       
/*
        // Set i to the index of the start item greater than ch
        // We know we will terminate without length test!
        int i = -1;
        while (true) {
            if (c < list[++i]) break;
        }
        */


       
int i = findCodePoint(c);

       
return ((i & 1) != 0); // return true if odd
   
}

   
/**
     * Returns the smallest value i such that c < list[i].  Caller
     * must ensure that c is a legal value or this method will enter
     * an infinite loop.  This method performs a binary search.
     * @param c a character in the range MIN_VALUE..MAX_VALUE
     * inclusive
     * @return the smallest integer i in the range 0..len-1,
     * inclusive, such that c < list[i]
     */

   
private final int findCodePoint(int c) {
       
/* Examples:
                                           findCodePoint(c)
           set              list[]         c=0 1 3 4 7 8
           ===              ==============   ===========
           []               [110000]         0 0 0 0 0 0
           [\u0000-\u0003]  [0, 4, 110000]   1 1 1 2 2 2
           [\u0004-\u0007]  [4, 8, 110000]   0 0 0 1 1 2
           [:all:]          [0, 110000]      1 1 1 1 1 1
         */


       
// Return the smallest i such that c < list[i].  Assume
       
// list[len - 1] == HIGH and that c is legal (0..HIGH-1).
       
if (c < list[0]) return 0;
       
// High runner test.  c is often after the last range, so an
       
// initial check for this condition pays off.
       
if (len >= 2 && c >= list[len-2]) return len-1;
       
int lo = 0;
       
int hi = len - 1;
       
// invariant: c >= list[lo]
       
// invariant: c < list[hi]
       
for (;;) {
           
int i = (lo + hi) >>> 1;
           
if (i == lo) return hi;
           
if (c < list[i]) {
                hi
= i;
           
} else {
                lo
= i;
           
}
       
}
   
}

   
/**
     * Adds all of the elements in the specified set to this set if
     * they're not already present.  This operation effectively
     * modifies this set so that its value is the <i>union</i> of the two
     * sets.  The behavior of this operation is unspecified if the specified
     * collection is modified while the operation is in progress.
     *
     * @param c set whose elements are to be added to this set.
     * @stable ICU 2.0
     */

   
public UnicodeSet addAll(UnicodeSet c) {
        add
(c.list, c.len, 0);
        strings
.addAll(c.strings);
       
return this;
   
}

   
/**
     * Retains only the elements in this set that are contained in the
     * specified set.  In other words, removes from this set all of
     * its elements that are not contained in the specified set.  This
     * operation effectively modifies this set so that its value is
     * the <i>intersection</i> of the two sets.
     *
     * @param c set that defines which elements this set will retain.
     * @stable ICU 2.0
     */

   
public UnicodeSet retainAll(UnicodeSet c) {
        retain
(c.list, c.len, 0);
        strings
.retainAll(c.strings);
       
return this;
   
}

   
/**
     * Removes from this set all of its elements that are contained in the
     * specified set.  This operation effectively modifies this
     * set so that its value is the <i>asymmetric set difference</i> of
     * the two sets.
     *
     * @param c set that defines which elements will be removed from
     *          this set.
     * @stable ICU 2.0
     */

   
public UnicodeSet removeAll(UnicodeSet c) {
        retain
(c.list, c.len, 2);
        strings
.removeAll(c.strings);
       
return this;
   
}

   
/**
     * Removes all of the elements from this set.  This set will be
     * empty after this call returns.
     * @stable ICU 2.0
     */

   
public UnicodeSet clear() {
        list
[0] = HIGH;
        len
= 1;
        pat
= null;
        strings
.clear();
       
return this;
   
}

   
/**
     * Iteration method that returns the number of ranges contained in
     * this set.
     * @see #getRangeStart
     * @see #getRangeEnd
     * @stable ICU 2.0
     */

   
public int getRangeCount() {
       
return len/2;
   
}

   
/**
     * Iteration method that returns the first character in the
     * specified range of this set.
     * @exception ArrayIndexOutOfBoundsException if index is outside
     * the range <code>0..getRangeCount()-1</code>
     * @see #getRangeCount
     * @see #getRangeEnd
     * @stable ICU 2.0
     */

   
public int getRangeStart(int index) {
       
return list[index*2];
   
}

   
/**
     * Iteration method that returns the last character in the
     * specified range of this set.
     * @exception ArrayIndexOutOfBoundsException if index is outside
     * the range <code>0..getRangeCount()-1</code>
     * @see #getRangeStart
     * @see #getRangeEnd
     * @stable ICU 2.0
     */

   
public int getRangeEnd(int index) {
       
return (list[index*2 + 1] - 1);
   
}

   
//----------------------------------------------------------------
   
// Implementation: Pattern parsing
   
//----------------------------------------------------------------

   
/**
     * Parses the given pattern, starting at the given position.  The character
     * at pattern.charAt(pos.getIndex()) must be '[', or the parse fails.
     * Parsing continues until the corresponding closing ']'.  If a syntax error
     * is encountered between the opening and closing brace, the parse fails.
     * Upon return from a successful parse, the ParsePosition is updated to
     * point to the character following the closing ']', and an inversion
     * list for the parsed pattern is returned.  This method
     * calls itself recursively to parse embedded subpatterns.
     *
     * @param pattern the string containing the pattern to be parsed.  The
     * portion of the string from pos.getIndex(), which must be a '[', to the
     * corresponding closing ']', is parsed.
     * @param pos upon entry, the position at which to being parsing.  The
     * character at pattern.charAt(pos.getIndex()) must be a '['.  Upon return
     * from a successful parse, pos.getIndex() is either the character after the
     * closing ']' of the parsed pattern, or pattern.length() if the closing ']'
     * is the last character of the pattern string.
     * @return an inversion list for the parsed substring
     * of <code>pattern</code>
     * @exception java.lang.IllegalArgumentException if the parse fails.
     */

   
UnicodeSet applyPattern(String pattern,
                     
ParsePosition pos,
                     
SymbolTable symbols,
                     
int options) {

       
// Need to build the pattern in a temporary string because
       
// _applyPattern calls add() etc., which set pat to empty.
       
boolean parsePositionWasNull = pos == null;
       
if (parsePositionWasNull) {
            pos
= new ParsePosition(0);
       
}

       
StringBuffer rebuiltPat = new StringBuffer();
       
RuleCharacterIterator chars =
           
new RuleCharacterIterator(pattern, symbols, pos);
        applyPattern
(chars, symbols, rebuiltPat, options);
       
if (chars.inVariable()) {
            syntaxError
(chars, "Extra chars in variable value");
       
}
        pat
= rebuiltPat.toString();
       
if (parsePositionWasNull) {
           
int i = pos.getIndex();

           
// Skip over trailing whitespace
           
if ((options & IGNORE_SPACE) != 0) {
                i
= Utility.skipWhitespace(pattern, i);
           
}

           
if (i != pattern.length()) {
               
throw new IllegalArgumentException("Parse of \"" + pattern +
                                                   
"\" failed at " + i);
           
}
       
}
       
return this;
   
}

   
/**
     * Parse the pattern from the given RuleCharacterIterator.  The
     * iterator is advanced over the parsed pattern.
     * @param chars iterator over the pattern characters.  Upon return
     * it will be advanced to the first character after the parsed
     * pattern, or the end of the iteration if all characters are
     * parsed.
     * @param symbols symbol table to use to parse and dereference
     * variables, or null if none.
     * @param rebuiltPat the pattern that was parsed, rebuilt or
     * copied from the input pattern, as appropriate.
     * @param options a bit mask of zero or more of the following:
     * IGNORE_SPACE, CASE.
     */

   
void applyPattern(RuleCharacterIterator chars, SymbolTable symbols,
                     
StringBuffer rebuiltPat, int options) {

       
// Syntax characters: [ ] ^ - & { }

       
// Recognized special forms for chars, sets: c-c s-s s&s

       
int opts = RuleCharacterIterator.PARSE_VARIABLES |
                   
RuleCharacterIterator.PARSE_ESCAPES;
       
if ((options & IGNORE_SPACE) != 0) {
            opts
|= RuleCharacterIterator.SKIP_WHITESPACE;
       
}

       
StringBuffer pat = new StringBuffer(), buf = null;
       
boolean usePat = false;
       
UnicodeSet scratch = null;
       
Object backup = null;

       
// mode: 0=before [, 1=between [...], 2=after ]
       
// lastItem: 0=none, 1=char, 2=set
       
int lastItem = 0, lastChar = 0, mode = 0;
       
char op = 0;

       
boolean invert = false;

        clear
();

       
while (mode != 2 && !chars.atEnd()) {
           
if (false) {
               
// Debugging assertion
               
if (!((lastItem == 0 && op == 0) ||
                     
(lastItem == 1 && (op == 0 || op == '-')) ||
                     
(lastItem == 2 && (op == 0 || op == '-' || op == '&')))) {
                   
throw new IllegalArgumentException();
               
}
           
}

           
int c = 0;
           
boolean literal = false;
           
UnicodeSet nested = null;

           
// -------- Check for property pattern

           
// setMode: 0=none, 1=unicodeset, 2=propertypat, 3=preparsed
           
int setMode = 0;
           
if (resemblesPropertyPattern(chars, opts)) {
                setMode
= 2;
           
}

           
// -------- Parse '[' of opening delimiter OR nested set.
           
// If there is a nested set, use `setMode' to define how
           
// the set should be parsed.  If the '[' is part of the
           
// opening delimiter for this pattern, parse special
           
// strings "[", "[^", "[-", and "[^-".  Check for stand-in
           
// characters representing a nested set in the symbol
           
// table.

           
else {
               
// Prepare to backup if necessary
                backup
= chars.getPos(backup);
                c
= chars.next(opts);
                literal
= chars.isEscaped();

               
if (c == '[' && !literal) {
                   
if (mode == 1) {
                        chars
.setPos(backup); // backup
                        setMode
= 1;
                   
} else {
                       
// Handle opening '[' delimiter
                        mode
= 1;
                        pat
.append('[');
                        backup
= chars.getPos(backup); // prepare to backup
                        c
= chars.next(opts);
                        literal
= chars.isEscaped();
                       
if (c == '^' && !literal) {
                            invert
= true;
                            pat
.append('^');
                            backup
= chars.getPos(backup); // prepare to backup
                            c
= chars.next(opts);
                            literal
= chars.isEscaped();
                       
}
                       
// Fall through to handle special leading '-';
                       
// otherwise restart loop for nested [], \p{}, etc.
                       
if (c == '-') {
                            literal
= true;
                           
// Fall through to handle literal '-' below
                       
} else {
                            chars
.setPos(backup); // backup
                           
continue;
                       
}
                   
}
               
} else if (symbols != null) {
                     
UnicodeMatcher m = symbols.lookupMatcher(c); // may be null
                     
if (m != null) {
                         
try {
                             nested
= (UnicodeSet) m;
                             setMode
= 3;
                         
} catch (ClassCastException e) {
                             syntaxError
(chars, "Syntax error");
                         
}
                     
}
               
}
           
}

           
// -------- Handle a nested set.  This either is inline in
           
// the pattern or represented by a stand-in that has
           
// previously been parsed and was looked up in the symbol
           
// table.

           
if (setMode != 0) {
               
if (lastItem == 1) {
                   
if (op != 0) {
                        syntaxError
(chars, "Char expected after operator");
                   
}
                    add
(lastChar, lastChar);
                    _appendToPat
(pat, lastChar, false);
                    lastItem
= op = 0;
               
}

               
if (op == '-' || op == '&') {
                    pat
.append(op);
               
}

               
if (nested == null) {
                   
if (scratch == null) scratch = new UnicodeSet();
                    nested
= scratch;
               
}
               
switch (setMode) {
               
case 1:
                    nested
.applyPattern(chars, symbols, pat, options);
                   
break;
               
case 2:
                    chars
.skipIgnored(opts);
                    nested
.applyPropertyPattern(chars, pat, symbols);
                   
break;
               
case 3: // `nested' already parsed
                    nested
._toPattern(pat, false);
                   
break;
               
}

                usePat
= true;

               
if (mode == 0) {
                   
// Entire pattern is a category; leave parse loop
                   
set(nested);
                    mode
= 2;
                   
break;
               
}

               
switch (op) {
               
case '-':
                    removeAll
(nested);
                   
break;
               
case '&':
                    retainAll
(nested);
                   
break;
               
case 0:
                    addAll
(nested);
                   
break;
               
}

                op
= 0;
                lastItem
= 2;

               
continue;
           
}

           
if (mode == 0) {
                syntaxError
(chars, "Missing '['");
           
}

           
// -------- Parse special (syntax) characters.  If the
           
// current character is not special, or if it is escaped,
           
// then fall through and handle it below.

           
if (!literal) {
               
switch (c) {
               
case ']':
                   
if (lastItem == 1) {
                        add
(lastChar, lastChar);
                        _appendToPat
(pat, lastChar, false);
                   
}
                   
// Treat final trailing '-' as a literal
                   
if (op == '-') {
                        add
(op, op);
                        pat
.append(op);
                   
} else if (op == '&') {
                        syntaxError
(chars, "Trailing '&'");
                   
}
                    pat
.append(']');
                    mode
= 2;
                   
continue;
               
case '-':
                   
if (op == 0) {
                       
if (lastItem != 0) {
                            op
= (char) c;
                           
continue;
                       
} else {
                           
// Treat final trailing '-' as a literal
                            add
(c, c);
                            c
= chars.next(opts);
                            literal
= chars.isEscaped();
                           
if (c == ']' && !literal) {
                                pat
.append("-]");
                                mode
= 2;
                               
continue;
                           
}
                       
}
                   
}
                    syntaxError
(chars, "'-' not after char or set");
               
case '&':
                   
if (lastItem == 2 && op == 0) {
                        op
= (char) c;
                       
continue;
                   
}
                    syntaxError
(chars, "'&' not after set");
               
case '^':
                    syntaxError
(chars, "'^' not after '['");
               
case '{':
                   
if (op != 0) {
                        syntaxError
(chars, "Missing operand after operator");
                   
}
                   
if (lastItem == 1) {
                        add
(lastChar, lastChar);
                        _appendToPat
(pat, lastChar, false);
                   
}
                    lastItem
= 0;
                   
if (buf == null) {
                        buf
= new StringBuffer();
                   
} else {
                        buf
.setLength(0);
                   
}
                   
boolean ok = false;
                   
while (!chars.atEnd()) {
                        c
= chars.next(opts);
                        literal
= chars.isEscaped();
                       
if (c == '}' && !literal) {
                            ok
= true;
                           
break;
                       
}
                        UTF16
.append(buf, c);
                   
}
                   
if (buf.length() < 1 || !ok) {
                        syntaxError
(chars, "Invalid multicharacter string");
                   
}
                   
// We have new string. Add it to set and continue;
                   
// we don't need to drop through to the further
                   
// processing
                    add
(buf.toString());
                    pat
.append('{');
                    _appendToPat
(pat, buf.toString(), false);
                    pat
.append('}');
                   
continue;
               
case SymbolTable.SYMBOL_REF:
                   
//         symbols  nosymbols
                   
// [a-$]   error    error (ambiguous)
                   
// [a$]    anchor   anchor
                   
// [a-$x]  var "x"* literal '$'
                   
// [a-$.]  error    literal '$'
                   
// *We won't get here in the case of var "x"
                    backup
= chars.getPos(backup);
                    c
= chars.next(opts);
                    literal
= chars.isEscaped();
                   
boolean anchor = (c == ']' && !literal);
                   
if (symbols == null && !anchor) {
                        c
= SymbolTable.SYMBOL_REF;
                        chars
.setPos(backup);
                       
break; // literal '$'
                   
}
                   
if (anchor && op == 0) {
                       
if (lastItem == 1) {
                            add
(lastChar, lastChar);
                            _appendToPat
(pat, lastChar, false);
                       
}
                        add
(UnicodeMatcher.ETHER);
                        usePat
= true;
                        pat
.append(SymbolTable.SYMBOL_REF).append(']');
                        mode
= 2;
                       
continue;
                   
}
                    syntaxError
(chars, "Unquoted '$'");
               
default:
                   
break;
               
}
           
}

           
// -------- Parse literal characters.  This includes both
           
// escaped chars ("\u4E01") and non-syntax characters
           
// ("a").

           
switch (lastItem) {
           
case 0:
                lastItem
= 1;
                lastChar
= c;
               
break;
           
case 1:
               
if (op == '-') {
                   
if (lastChar >= c) {
                       
// Don't allow redundant (a-a) or empty (b-a) ranges;
                       
// these are most likely typos.
                        syntaxError
(chars, "Invalid range");
                   
}
                    add
(lastChar, c);
                    _appendToPat
(pat, lastChar, false);
                    pat
.append(op);
                    _appendToPat
(pat, c, false);
                    lastItem
= op = 0;
               
} else {
                    add
(lastChar, lastChar);
                    _appendToPat
(pat, lastChar, false);
                    lastChar
= c;
               
}
               
break;
           
case 2:
               
if (op != 0) {
                    syntaxError
(chars, "Set expected after operator");
               
}
                lastChar
= c;
                lastItem
= 1;
               
break;
           
}
       
}

       
if (mode != 2) {
            syntaxError
(chars, "Missing ']'");
       
}

        chars
.skipIgnored(opts);

       
if (invert) {
            complement
();
       
}

       
// Use the rebuilt pattern (pat) only if necessary.  Prefer the
       
// generated pattern.
       
if (usePat) {
            rebuiltPat
.append(pat.toString());
       
} else {
            _generatePattern
(rebuiltPat, false);
       
}
   
}

   
private static void syntaxError(RuleCharacterIterator chars, String msg) {
       
throw new IllegalArgumentException("Error: " + msg + " at \"" +
                                           
Utility.escape(chars.toString()) +
                                           
'"');
   
}

   
//----------------------------------------------------------------
   
// Implementation: Utility methods
   
//----------------------------------------------------------------

   
private void ensureCapacity(int newLen) {
       
if (newLen <= list.length) return;
       
int[] temp = new int[newLen + GROW_EXTRA];
       
System.arraycopy(list, 0, temp, 0, len);
        list
= temp;
   
}

   
private void ensureBufferCapacity(int newLen) {
       
if (buffer != null && newLen <= buffer.length) return;
        buffer
= new int[newLen + GROW_EXTRA];
   
}

   
/**
     * Assumes start <= end.
     */

   
private int[] range(int start, int end) {
       
if (rangeList == null) {
            rangeList
= new int[] { start, end+1, HIGH };
       
} else {
            rangeList
[0] = start;
            rangeList
[1] = end+1;
       
}
       
return rangeList;
   
}

   
//----------------------------------------------------------------
   
// Implementation: Fundamental operations
   
//----------------------------------------------------------------

   
// polarity = 0, 3 is normal: x xor y
   
// polarity = 1, 2: x xor ~y == x === y

   
private UnicodeSet xor(int[] other, int otherLen, int polarity) {
        ensureBufferCapacity
(len + otherLen);
       
int i = 0, j = 0, k = 0;
       
int a = list[i++];
       
int b;
       
if (polarity == 1 || polarity == 2) {
            b
= LOW;
           
if (other[j] == LOW) { // skip base if already LOW
               
++j;
                b
= other[j];
           
}
       
} else {
            b
= other[j++];
       
}
       
// simplest of all the routines
       
// sort the values, discarding identicals!
       
while (true) {
           
if (a < b) {
                buffer
[k++] = a;
                a
= list[i++];
           
} else if (b < a) {
                buffer
[k++] = b;
                b
= other[j++];
           
} else if (a != HIGH) { // at this point, a == b
               
// discard both values!
                a
= list[i++];
                b
= other[j++];
           
} else { // DONE!
                buffer
[k++] = HIGH;
                len
= k;
               
break;
           
}
       
}
       
// swap list and buffer
       
int[] temp = list;
        list
= buffer;
        buffer
= temp;
        pat
= null;
       
return this;
   
}

   
// polarity = 0 is normal: x union y
   
// polarity = 2: x union ~y
   
// polarity = 1: ~x union y
   
// polarity = 3: ~x union ~y

   
private UnicodeSet add(int[] other, int otherLen, int polarity) {
        ensureBufferCapacity
(len + otherLen);
       
int i = 0, j = 0, k = 0;
       
int a = list[i++];
       
int b = other[j++];
       
// change from xor is that we have to check overlapping pairs
       
// polarity bit 1 means a is second, bit 2 means b is.
        main
:
       
while (true) {
           
switch (polarity) {
             
case 0: // both first; take lower if unequal
               
if (a < b) { // take a
                   
// Back up over overlapping ranges in buffer[]
                   
if (k > 0 && a <= buffer[k-1]) {
                       
// Pick latter end value in buffer[] vs. list[]
                        a
= max(list[i], buffer[--k]);
                   
} else {
                       
// No overlap
                        buffer
[k++] = a;
                        a
= list[i];
                   
}
                    i
++; // Common if/else code factored out
                    polarity
^= 1;
               
} else if (b < a) { // take b
                   
if (k > 0 && b <= buffer[k-1]) {
                        b
= max(other[j], buffer[--k]);
                   
} else {
                        buffer
[k++] = b;
                        b
= other[j];
                   
}
                    j
++;
                    polarity
^= 2;
               
} else { // a == b, take a, drop b
                   
if (a == HIGH) break main;
                   
// This is symmetrical; it doesn't matter if
                   
// we backtrack with a or b. - liu
                   
if (k > 0 && a <= buffer[k-1]) {
                        a
= max(list[i], buffer[--k]);
                   
} else {
                       
// No overlap
                        buffer
[k++] = a;
                        a
= list[i];
                   
}
                    i
++;
                    polarity
^= 1;
                    b
= other[j++]; polarity ^= 2;
               
}
               
break;
             
case 3: // both second; take higher if unequal, and drop other
               
if (b <= a) { // take a
                   
if (a == HIGH) break main;
                    buffer
[k++] = a;
               
} else { // take b
                   
if (b == HIGH) break main;
                    buffer
[k++] = b;
               
}
                a
= list[i++]; polarity ^= 1;   // factored common code
                b
= other[j++]; polarity ^= 2;
               
break;
             
case 1: // a second, b first; if b < a, overlap
               
if (a < b) { // no overlap, take a
                    buffer
[k++] = a; a = list[i++]; polarity ^= 1;
               
} else if (b < a) { // OVERLAP, drop b
                    b
= other[j++]; polarity ^= 2;
               
} else { // a == b, drop both!
                   
if (a == HIGH) break main;
                    a
= list[i++]; polarity ^= 1;
                    b
= other[j++]; polarity ^= 2;
               
}
               
break;
             
case 2: // a first, b second; if a < b, overlap
               
if (b < a) { // no overlap, take b
                    buffer
[k++] = b; b = other[j++]; polarity ^= 2;
               
} else  if (a < b) { // OVERLAP, drop a
                    a
= list[i++]; polarity ^= 1;
               
} else { // a == b, drop both!
                   
if (a == HIGH) break main;
                    a
= list[i++]; polarity ^= 1;
                    b
= other[j++]; polarity ^= 2;
               
}
               
break;
           
}
       
}
        buffer
[k++] = HIGH;    // terminate
        len
= k;
       
// swap list and buffer
       
int[] temp = list;
        list
= buffer;
        buffer
= temp;
        pat
= null;
       
return this;
   
}

   
// polarity = 0 is normal: x intersect y
   
// polarity = 2: x intersect ~y == set-minus
   
// polarity = 1: ~x intersect y
   
// polarity = 3: ~x intersect ~y

   
private UnicodeSet retain(int[] other, int otherLen, int polarity) {
        ensureBufferCapacity
(len + otherLen);
       
int i = 0, j = 0, k = 0;
       
int a = list[i++];
       
int b = other[j++];
       
// change from xor is that we have to check overlapping pairs
       
// polarity bit 1 means a is second, bit 2 means b is.
        main
:
       
while (true) {
           
switch (polarity) {
             
case 0: // both first; drop the smaller
               
if (a < b) { // drop a
                    a
= list[i++]; polarity ^= 1;
               
} else if (b < a) { // drop b
                    b
= other[j++]; polarity ^= 2;
               
} else { // a == b, take one, drop other
                   
if (a == HIGH) break main;
                    buffer
[k++] = a; a = list[i++]; polarity ^= 1;
                    b
= other[j++]; polarity ^= 2;
               
}
               
break;
             
case 3: // both second; take lower if unequal
               
if (a < b) { // take a
                    buffer
[k++] = a; a = list[i++]; polarity ^= 1;
               
} else if (b < a) { // take b
                    buffer
[k++] = b; b = other[j++]; polarity ^= 2;
               
} else { // a == b, take one, drop other
                   
if (a == HIGH) break main;
                    buffer
[k++] = a; a = list[i++]; polarity ^= 1;
                    b
= other[j++]; polarity ^= 2;
               
}
               
break;
             
case 1: // a second, b first;
               
if (a < b) { // NO OVERLAP, drop a
                    a
= list[i++]; polarity ^= 1;
               
} else if (b < a) { // OVERLAP, take b
                    buffer
[k++] = b; b = other[j++]; polarity ^= 2;
               
} else { // a == b, drop both!
                   
if (a == HIGH) break main;
                    a
= list[i++]; polarity ^= 1;
                    b
= other[j++]; polarity ^= 2;
               
}
               
break;
             
case 2: // a first, b second; if a < b, overlap
               
if (b < a) { // no overlap, drop b
                    b
= other[j++]; polarity ^= 2;
               
} else  if (a < b) { // OVERLAP, take a
                    buffer
[k++] = a; a = list[i++]; polarity ^= 1;
               
} else { // a == b, drop both!
                   
if (a == HIGH) break main;
                    a
= list[i++]; polarity ^= 1;
                    b
= other[j++]; polarity ^= 2;
               
}
               
break;
           
}
       
}
        buffer
[k++] = HIGH;    // terminate
        len
= k;
       
// swap list and buffer
       
int[] temp = list;
        list
= buffer;
        buffer
= temp;
        pat
= null;
       
return this;
   
}

   
private static final int max(int a, int b) {
       
return (a > b) ? a : b;
   
}

   
//----------------------------------------------------------------
   
// Generic filter-based scanning code
   
//----------------------------------------------------------------

   
private static interface Filter {
       
boolean contains(int codePoint);
   
}

   
// VersionInfo for unassigned characters
   
static final VersionInfo NO_VERSION = VersionInfo.getInstance(0, 0, 0, 0);

   
private static class VersionFilter implements Filter {
       
VersionInfo version;
       
VersionFilter(VersionInfo version) { this.version = version; }
       
public boolean contains(int ch) {
           
VersionInfo v = UCharacter.getAge(ch);
           
// Reference comparison ok; VersionInfo caches and reuses
           
// unique objects.
           
return v != NO_VERSION &&
                   v
.compareTo(version) <= 0;
       
}
   
}

   
private static synchronized UnicodeSet getInclusions() {
       
if (INCLUSIONS == null) {
           
UCharacterProperty property = UCharacterProperty.getInstance();
            INCLUSIONS
= property.getInclusions();
       
}
       
return INCLUSIONS;
   
}

   
/**
     * Generic filter-based scanning code for UCD property UnicodeSets.
     */

   
private UnicodeSet applyFilter(Filter filter) {
       
// Walk through all Unicode characters, noting the start
       
// and end of each range for which filter.contain(c) is
       
// true.  Add each range to a set.
       
//
       
// To improve performance, use the INCLUSIONS set, which
       
// encodes information about character ranges that are known
       
// to have identical properties, such as the CJK Ideographs
       
// from U+4E00 to U+9FA5.  INCLUSIONS contains all characters
       
// except the first characters of such ranges.
       
//
       
// TODO Where possible, instead of scanning over code points,
       
// use internal property data to initialize UnicodeSets for
       
// those properties.  Scanning code points is slow.

        clear
();

       
int startHasProperty = -1;
       
UnicodeSet inclusions = getInclusions();
       
int limitRange = inclusions.getRangeCount();

       
for (int j=0; j<limitRange; ++j) {
           
// get current range
           
int start = inclusions.getRangeStart(j);
           
int end = inclusions.getRangeEnd(j);

           
// for all the code points in the range, process
           
for (int ch = start; ch <= end; ++ch) {
               
// only add to the unicodeset on inflection points --
               
// where the hasProperty value changes to false
               
if (filter.contains(ch)) {
                   
if (startHasProperty < 0) {
                        startHasProperty
= ch;
                   
}
               
} else if (startHasProperty >= 0) {
                    add
(startHasProperty, ch-1);
                    startHasProperty
= -1;
               
}
           
}
       
}
       
if (startHasProperty >= 0) {
            add
(startHasProperty, 0x10FFFF);
       
}

       
return this;
   
}


   
/**
     * Remove leading and trailing rule white space and compress
     * internal rule white space to a single space character.
     *
     * @see UCharacterProperty#isRuleWhiteSpace
     */

   
private static String mungeCharName(String source) {
       
StringBuffer buf = new StringBuffer();
       
for (int i=0; i<source.length(); ) {
           
int ch = UTF16.charAt(source, i);
            i
+= UTF16.getCharCount(ch);
           
if (UCharacterProperty.isRuleWhiteSpace(ch)) {
               
if (buf.length() == 0 ||
                    buf
.charAt(buf.length() - 1) == ' ') {
                   
continue;
               
}
                ch
= ' '; // convert to ' '
           
}
            UTF16
.append(buf, ch);
       
}
       
if (buf.length() != 0 &&
            buf
.charAt(buf.length() - 1) == ' ') {
            buf
.setLength(buf.length() - 1);
       
}
       
return buf.toString();
   
}

   
//----------------------------------------------------------------
   
// Property set API
   
//----------------------------------------------------------------

   
/**
     * Modifies this set to contain those code points which have the
     * given value for the given property.  Prior contents of this
     * set are lost.
     * @param propertyAlias
     * @param valueAlias
     * @param symbols if not null, then symbols are first called to see if a property
     * is available. If true, then everything else is skipped.
     * @return this set
     * @draft ICU 3.2
     * @deprecated This is a draft API and might change in a future release of ICU.
     */

   
public UnicodeSet applyPropertyAlias(String propertyAlias,
                                         
String valueAlias, SymbolTable symbols) {
               
if (propertyAlias.equals("Age"))
                   
{
                       
// Must munge name, since
                       
// VersionInfo.getInstance() does not do
                       
// 'loose' matching.
                       
VersionInfo version = VersionInfo.getInstance(mungeCharName(valueAlias));
                        applyFilter
(new VersionFilter(version));
                       
return this;
                   
}
               
else
                   
throw new IllegalArgumentException("Unsupported property");
   
}

   
/**
     * Return true if the given iterator appears to point at a
     * property pattern.  Regardless of the result, return with the
     * iterator unchanged.
     * @param chars iterator over the pattern characters.  Upon return
     * it will be unchanged.
     * @param iterOpts RuleCharacterIterator options
     */

   
private static boolean resemblesPropertyPattern(RuleCharacterIterator chars,
                                                   
int iterOpts) {
       
boolean result = false;
        iterOpts
&= ~RuleCharacterIterator.PARSE_ESCAPES;
       
Object pos = chars.getPos(null);
       
int c = chars.next(iterOpts);
       
if (c == '[' || c == '\\') {
           
int d = chars.next(iterOpts & ~RuleCharacterIterator.SKIP_WHITESPACE);
            result
= (c == '[') ? (d == ':') :
                     
(d == 'N' || d == 'p' || d == 'P');
       
}
        chars
.setPos(pos);
       
return result;
   
}

   
/**
     * Parse the given property pattern at the given parse position.
     * @param symbols TODO
     */

   
private UnicodeSet applyPropertyPattern(String pattern, ParsePosition ppos, SymbolTable symbols) {
       
int pos = ppos.getIndex();

       
// On entry, ppos should point to one of the following locations:

       
// Minimum length is 5 characters, e.g. \p{L}
       
if ((pos+5) > pattern.length()) {
           
return null;
       
}

       
boolean posix = false; // true for [:pat:], false for \p{pat} \P{pat} \N{pat}
       
boolean isName = false; // true for \N{pat}, o/w false
       
boolean invert = false;

       
// Look for an opening [:, [:^, \p, or \P
       
if (pattern.regionMatches(pos, "[:", 0, 2)) {
            posix
= true;
            pos
= Utility.skipWhitespace(pattern, pos+2);
           
if (pos < pattern.length() && pattern.charAt(pos) == '^') {
               
++pos;
                invert
= true;
           
}
       
} else if (pattern.regionMatches(true, pos, "\\p", 0, 2) ||
                   pattern
.regionMatches(pos, "\\N", 0, 2)) {
           
char c = pattern.charAt(pos+1);
            invert
= (c == 'P');
            isName
= (c == 'N');
            pos
= Utility.skipWhitespace(pattern, pos+2);
           
if (pos == pattern.length() || pattern.charAt(pos++) != '{') {
               
// Syntax error; "\p" or "\P" not followed by "{"
               
return null;
           
}
       
} else {
           
// Open delimiter not seen
           
return null;
       
}

       
// Look for the matching close delimiter, either :] or }
       
int close = pattern.indexOf(posix ? ":]" : "}", pos);
       
if (close < 0) {
           
// Syntax error; close delimiter missing
           
return null;
       
}

       
// Look for an '=' sign.  If this is present, we will parse a
       
// medium \p{gc=Cf} or long \p{GeneralCategory=Format}
       
// pattern.
       
int equals = pattern.indexOf('=', pos);
       
String propName, valueName;
       
if (equals >= 0 && equals < close && !isName) {
           
// Equals seen; parse medium/long pattern
            propName
= pattern.substring(pos, equals);
            valueName
= pattern.substring(equals+1, close);
       
}

       
else {
           
// Handle case where no '=' is seen, and \N{}
            propName
= pattern.substring(pos, close);
            valueName
= "";

           
// Handle \N{name}
           
if (isName) {
               
// This is a little inefficient since it means we have to
               
// parse "na" back to UProperty.NAME even though we already
               
// know it's UProperty.NAME.  If we refactor the API to
               
// support args of (int, String) then we can remove
               
// "na" and make this a little more efficient.
                valueName
= propName;
                propName
= "na";
           
}
       
}

        applyPropertyAlias
(propName, valueName, symbols);

       
if (invert) {
            complement
();
       
}

       
// Move to the limit position after the close delimiter
        ppos
.setIndex(close + (posix ? 2 : 1));

       
return this;
   
}

   
/**
     * Parse a property pattern.
     * @param chars iterator over the pattern characters.  Upon return
     * it will be advanced to the first character after the parsed
     * pattern, or the end of the iteration if all characters are
     * parsed.
     * @param rebuiltPat the pattern that was parsed, rebuilt or
     * copied from the input pattern, as appropriate.
     * @param symbols TODO
     */

   
private void applyPropertyPattern(RuleCharacterIterator chars,
                                     
StringBuffer rebuiltPat, SymbolTable symbols) {
       
String pat = chars.lookahead();
       
ParsePosition pos = new ParsePosition(0);
        applyPropertyPattern
(pat, pos, symbols);
       
if (pos.getIndex() == 0) {
            syntaxError
(chars, "Invalid property pattern");
       
}
        chars
.jumpahead(pos.getIndex());
        rebuiltPat
.append(pat.substring(0, pos.getIndex()));
   
}

   
//----------------------------------------------------------------
   
// Case folding API
   
//----------------------------------------------------------------

   
/**
     * Bitmask for constructor and applyPattern() indicating that
     * white space should be ignored.  If set, ignore characters for
     * which UCharacterProperty.isRuleWhiteSpace() returns true,
     * unless they are quoted or escaped.  This may be ORed together
     * with other selectors.
     * @internal
     */

   
public static final int IGNORE_SPACE = 1;

}