/*

 * Copyright (c) 1997, 2019, 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 Taligent, Inc. 1996, 1997 - All Rights Reserved

 * (C) Copyright IBM Corp. 1996-1998 - All Rights Reserved

 *

 *   The original version of this source code and documentation is copyrighted

 * and owned by Taligent, Inc., a wholly-owned subsidiary of IBM. These

 * materials are provided under terms of a License Agreement between Taligent

 * and Sun. This technology is protected by multiple US and International

 * patents. This notice and attribution to Taligent may not be removed.

 *   Taligent is a registered trademark of Taligent, Inc.

 *

 */

package java.text;

import java.text.Normalizer;

import java.util.Vector;

import java.util.Locale;

/**

 * The {@code RuleBasedCollator} class is a concrete subclass of

 * {@code Collator} that provides a simple, data-driven, table

 * collator.  With this class you can create a customized table-based

 * {@code Collator}.  {@code RuleBasedCollator} maps

 * characters to sort keys.

 *

 * <p>

 * {@code RuleBasedCollator} has the following restrictions

 * for efficiency (other subclasses may be used for more complex languages) :

 * <ol>

 * <li>If a special collation rule controlled by a &lt;modifier&gt; is

      specified it applies to the whole collator object.

 * <li>All non-mentioned characters are at the end of the

 *     collation order.

 * </ol>

 *

 * <p>

 * The collation table is composed of a list of collation rules, where each

 * rule is of one of three forms:

 * <pre>

 *    <modifier>

 *    <relation> <text-argument>

 *    <reset> <text-argument>

 * </pre>

 * The definitions of the rule elements is as follows:

 * <UL>

 *    <LI><strong>Text-Argument</strong>: A text-argument is any sequence of

 *        characters, excluding special characters (that is, common

 *        whitespace characters [0009-000D, 0020] and rule syntax characters

 *        [0021-002F, 003A-0040, 005B-0060, 007B-007E]). If those

 *        characters are desired, you can put them in single quotes

 *        (e.g. ampersand => '&'). Note that unquoted white space characters

 *        are ignored; e.g. {@code b c} is treated as {@code bc}.

 *    <LI><strong>Modifier</strong>: There are currently two modifiers that

 *        turn on special collation rules.

 *        <UL>

 *            <LI>'@' : Turns on backwards sorting of accents (secondary

 *                      differences), as in French.

 *            <LI>'!' : Turns on Thai/Lao vowel-consonant swapping.  If this

 *                      rule is in force when a Thai vowel of the range

 *                      \U0E40-\U0E44 precedes a Thai consonant of the range

 *                      \U0E01-\U0E2E OR a Lao vowel of the range \U0EC0-\U0EC4

 *                      precedes a Lao consonant of the range \U0E81-\U0EAE then

 *                      the vowel is placed after the consonant for collation

 *                      purposes.

 *        </UL>

 *        <p>'@' : Indicates that accents are sorted backwards, as in French.

 *    <LI><strong>Relation</strong>: The relations are the following:

 *        <UL>

 *            <LI>'&lt;' : Greater, as a letter difference (primary)

 *            <LI>';' : Greater, as an accent difference (secondary)

 *            <LI>',' : Greater, as a case difference (tertiary)

 *            <LI>'=' : Equal

 *        </UL>

 *    <LI><strong>Reset</strong>: There is a single reset

 *        which is used primarily for contractions and expansions, but which

 *        can also be used to add a modification at the end of a set of rules.

 *        <p>'&amp;' : Indicates that the next rule follows the position to where

 *            the reset text-argument would be sorted.

 * </UL>

 *

 * <p>

 * This sounds more complicated than it is in practice. For example, the

 * following are equivalent ways of expressing the same thing:

 * <blockquote>

 * <pre>

 * a < b < c

 * a < b & b < c

 * a < c & a < b

 * </pre>

 * </blockquote>

 * Notice that the order is important, as the subsequent item goes immediately

 * after the text-argument. The following are not equivalent:

 * <blockquote>

 * <pre>

 * a < b & a < c

 * a < c & a < b

 * </pre>

 * </blockquote>

 * Either the text-argument must already be present in the sequence, or some

 * initial substring of the text-argument must be present. (e.g. "a < b & ae <

 * e" is valid since "a" is present in the sequence before "ae" is reset). In

 * this latter case, "ae" is not entered and treated as a single character;

 * instead, "e" is sorted as if it were expanded to two characters: "a"

 * followed by an "e". This difference appears in natural languages: in

 * traditional Spanish "ch" is treated as though it contracts to a single

 * character (expressed as "c < ch < d"), while in traditional German

 * a-umlaut is treated as though it expanded to two characters

 * (expressed as "a,A < b,B ... &ae;\u00e3&AE;\u00c3").

 * [\u00e3 and \u00c3 are, of course, the escape sequences for a-umlaut.]

 * <p>

 * <strong>Ignorable Characters</strong>

 * <p>

 * For ignorable characters, the first rule must start with a relation (the

 * examples we have used above are really fragments; "a < b" really should be

 * "< a < b"). If, however, the first relation is not "<", then all the all

 * text-arguments up to the first "<" are ignorable. For example, ", - < a < b"

 * makes "-" an ignorable character, as we saw earlier in the word

 * "black-birds". In the samples for different languages, you see that most

 * accents are ignorable.

 *

 * <p><strong>Normalization and Accents</strong>

 * <p>

 * {@code RuleBasedCollator} automatically processes its rule table to

 * include both pre-composed and combining-character versions of

 * accented characters.  Even if the provided rule string contains only

 * base characters and separate combining accent characters, the pre-composed

 * accented characters matching all canonical combinations of characters from

 * the rule string will be entered in the table.

 * <p>

 * This allows you to use a RuleBasedCollator to compare accented strings

 * even when the collator is set to NO_DECOMPOSITION.  There are two caveats,

 * however.  First, if the strings to be collated contain combining

 * sequences that may not be in canonical order, you should set the collator to

 * CANONICAL_DECOMPOSITION or FULL_DECOMPOSITION to enable sorting of

 * combining sequences.  Second, if the strings contain characters with

 * compatibility decompositions (such as full-width and half-width forms),

 * you must use FULL_DECOMPOSITION, since the rule tables only include

 * canonical mappings.

 *

 * <p><strong>Errors</strong>

 * <p>

 * The following are errors:

 * <UL>

 *     <LI>A text-argument contains unquoted punctuation symbols

 *        (e.g. "a < b-c < d").

 *     <LI>A relation or reset character not followed by a text-argument

 *        (e.g. "a < ,b").

 *     <LI>A reset where the text-argument (or an initial substring of the

 *         text-argument) is not already in the sequence.

 *         (e.g. "a < b & e < f")

 * </UL>

 * If you produce one of these errors, a {@code RuleBasedCollator} throws

 * a {@code ParseException}.

 *

 * <p><strong>Examples</strong>

 * <p>Simple:     "&lt; a &lt; b &lt; c &lt; d"

 * <p>Norwegian:  "&lt; a, A &lt; b, B &lt; c, C &lt; d, D &lt; e, E &lt; f, F

 *                 < g, G < h, H < i, I < j, J < k, K < l, L

 *                 < m, M < n, N < o, O < p, P < q, Q < r, R

 *                 < s, S < t, T < u, U < v, V < w, W < x, X

 *                 < y, Y < z, Z

 *                 < \u00E6, \u00C6

 *                 < \u00F8, \u00D8

 *                 < \u00E5 = a\u030A, \u00C5 = A\u030A;

 *                      aa, AA"

 *

 * <p>

 * To create a {@code RuleBasedCollator} object with specialized

 * rules tailored to your needs, you construct the {@code RuleBasedCollator}

 * with the rules contained in a {@code String} object. For example:

 * <blockquote>

 * <pre>

 * String simple = "< a< b< c< d";

 * RuleBasedCollator mySimple = new RuleBasedCollator(simple);

 * </pre>

 * </blockquote>

 * Or:

 * <blockquote>

 * <pre>

 * String Norwegian = "< a, A < b, B < c, C < d, D < e, E < f, F < g, G < h, H < i, I" +

 *                    "< j, J < k, K < l, L < m, M < n, N < o, O < p, P < q, Q < r, R" +

 *                    "< s, S < t, T < u, U < v, V < w, W < x, X < y, Y < z, Z" +

 *                    "< \u00E6, \u00C6" +     // Latin letter ae & AE

 *                    "< \u00F8, \u00D8" +     // Latin letter o & O with stroke

 *                    "< \u00E5 = a\u030A," +  // Latin letter a with ring above

 *                    "  \u00C5 = A\u030A;" +  // Latin letter A with ring above

 *                    "  aa, AA";

 * RuleBasedCollator myNorwegian = new RuleBasedCollator(Norwegian);

 * </pre>

 * </blockquote>

 *

 * <p>

 * A new collation rules string can be created by concatenating rules

 * strings. For example, the rules returned by {@link #getRules()} could

 * be concatenated to combine multiple {@code RuleBasedCollator}s.

 *

 * <p>

 * The following example demonstrates how to change the order of

 * non-spacing accents,

 * <blockquote>

 * <pre>

 * // old rule

 * String oldRules = "=\u0301;\u0300;\u0302;\u0308"    // main accents

 *                 + ";\u0327;\u0303;\u0304;\u0305"    // main accents

 *                 + ";\u0306;\u0307;\u0309;\u030A"    // main accents

 *                 + ";\u030B;\u030C;\u030D;\u030E"    // main accents

 *                 + ";\u030F;\u0310;\u0311;\u0312"    // main accents

 *                 + "< a , A ; ae, AE ; \u00e6 , \u00c6"

 *                 + "< b , B < c, C < e, E & C < d, D";

 * // change the order of accent characters

 * String addOn = "& \u0300 ; \u0308 ; \u0302";

 * RuleBasedCollator myCollator = new RuleBasedCollator(oldRules + addOn);

 * </pre>

 * </blockquote>

 *

 * @see        Collator

 * @see        CollationElementIterator

 * @author     Helena Shih, Laura Werner, Richard Gillam

 * @since 1.1

 */

public class RuleBasedCollator extends Collator{

    // IMPLEMENTATION NOTES:  The implementation of the collation algorithm is

    // divided across three classes: RuleBasedCollator, RBCollationTables, and

    // CollationElementIterator.  RuleBasedCollator contains the collator's

    // transient state and includes the code that uses the other classes to

    // implement comparison and sort-key building.  RuleBasedCollator also

    // contains the logic to handle French secondary accent sorting.

    // A RuleBasedCollator has two CollationElementIterators.  State doesn't

    // need to be preserved in these objects between calls to compare() or

    // getCollationKey(), but the objects persist anyway to avoid wasting extra

    // creation time.  compare() and getCollationKey() are synchronized to ensure

    // thread safety with this scheme.  The CollationElementIterator is responsible

    // for generating collation elements from strings and returning one element at

    // a time (sometimes there's a one-to-many or many-to-one mapping between

    // characters and collation elements-- this class handles that).

    // CollationElementIterator depends on RBCollationTables, which contains the

    // collator's static state.  RBCollationTables contains the actual data

    // tables specifying the collation order of characters for a particular locale

    // or use.  It also contains the base logic that CollationElementIterator

    // uses to map from characters to collation elements.  A single RBCollationTables

    // object is shared among all RuleBasedCollators for the same locale, and

    // thus by all the CollationElementIterators they create.

    /**

     * RuleBasedCollator constructor.  This takes the table rules and builds

     * a collation table out of them.  Please see RuleBasedCollator class

     * description for more details on the collation rule syntax.

     * @see java.util.Locale

     * @param rules the collation rules to build the collation table from.

     * @throws    ParseException A format exception

     * will be thrown if the build process of the rules fails. For

     * example, build rule "a < ? < d" will cause the constructor to

     * throw the ParseException because the '?' is not quoted.

     */

    public RuleBasedCollator(String rules) throws ParseException {

        this(rules, Collator.CANONICAL_DECOMPOSITION);

    }

    /**

     * RuleBasedCollator constructor.  This takes the table rules and builds

     * a collation table out of them.  Please see RuleBasedCollator class

     * description for more details on the collation rule syntax.

     * @see java.util.Locale

     * @param rules the collation rules to build the collation table from.

     * @param decomp the decomposition strength used to build the

     * collation table and to perform comparisons.

     * @throws    ParseException A format exception

     * will be thrown if the build process of the rules fails. For

     * example, build rule "a < ? < d" will cause the constructor to

     * throw the ParseException because the '?' is not quoted.

     */

    RuleBasedCollator(String rules, int decomp) throws ParseException {

        setStrength(Collator.TERTIARY);

        setDecomposition(decomp);

        tables = new RBCollationTables(rules, decomp);

    }

    /**

     * "Copy constructor."  Used in clone() for performance.

     */

    private RuleBasedCollator(RuleBasedCollator that) {

        setStrength(that.getStrength());

        setDecomposition(that.getDecomposition());

        tables = that.tables;

    }

    /**

     * Gets the table-based rules for the collation object.

     * @return returns the collation rules that the table collation object

     * was created from.

     */

    public String getRules()

    {

        return tables.getRules();

    }

    /**

     * Returns a CollationElementIterator for the given String.

     *

     * @param source the string to be collated

     * @return a {@code CollationElementIterator} object

     * @see java.text.CollationElementIterator

     */

    public CollationElementIterator getCollationElementIterator(String source) {

        return new CollationElementIterator( source, this );

    }

    /**

     * Returns a CollationElementIterator for the given CharacterIterator.

     *

     * @param source the character iterator to be collated

     * @return a {@code CollationElementIterator} object

     * @see java.text.CollationElementIterator

     * @since 1.2

     */

    public CollationElementIterator getCollationElementIterator(

                                                CharacterIterator source) {

        return new CollationElementIterator( source, this );

    }

    /**

     * Compares the character data stored in two different strings based on the

     * collation rules.  Returns information about whether a string is less

     * than, greater than or equal to another string in a language.

     * This can be overridden in a subclass.

     *

     * @throws    NullPointerException if {@code source} or {@code target} is null.

     */

    public synchronized int compare(String source, String target)

    {

        if (source == null || target == null) {

            throw new NullPointerException();

        }

        // The basic algorithm here is that we use CollationElementIterators

        // to step through both the source and target strings.  We compare each

        // collation element in the source string against the corresponding one

        // in the target, checking for differences.

        //

        // If a difference is found, we set <result> to LESS or GREATER to

        // indicate whether the source string is less or greater than the target.

        //

        // However, it's not that simple.  If we find a tertiary difference

        // (e.g. 'A' vs. 'a') near the beginning of a string, it can be

        // overridden by a primary difference (e.g. "A" vs. "B") later in

        // the string.  For example, "AA" < "aB", even though 'A' > 'a'.

        //

        // To keep track of this, we use strengthResult to keep track of the

        // strength of the most significant difference that has been found

        // so far.  When we find a difference whose strength is greater than

        // strengthResult, it overrides the last difference (if any) that

        // was found.

        int result = Collator.EQUAL;

        if (sourceCursor == null) {

            sourceCursor = getCollationElementIterator(source);

        } else {

            sourceCursor.setText(source);

        }

        if (targetCursor == null) {

            targetCursor = getCollationElementIterator(target);

        } else {

            targetCursor.setText(target);

        }

        int sOrder = 0, tOrder = 0;

        boolean initialCheckSecTer = getStrength() >= Collator.SECONDARY;

        boolean checkSecTer = initialCheckSecTer;

        boolean checkTertiary = getStrength() >= Collator.TERTIARY;

        boolean gets = true, gett = true;

        while(true) {

            // Get the next collation element in each of the strings, unless

            // we've been requested to skip it.

            if (gets) sOrder = sourceCursor.next(); else gets = true;

            if (gett) tOrder = targetCursor.next(); else gett = true;

            // If we've hit the end of one of the strings, jump out of the loop

            if ((sOrder == CollationElementIterator.NULLORDER)||

                (tOrder == CollationElementIterator.NULLORDER))

                break;

            int pSOrder = CollationElementIterator.primaryOrder(sOrder);

            int pTOrder = CollationElementIterator.primaryOrder(tOrder);

            // If there's no difference at this position, we can skip it

            if (sOrder == tOrder) {

                if (tables.isFrenchSec() && pSOrder != 0) {

                    if (!checkSecTer) {

                        // in french, a secondary difference more to the right is stronger,

                        // so accents have to be checked with each base element

                        checkSecTer = initialCheckSecTer;

                        // but tertiary differences are less important than the first

                        // secondary difference, so checking tertiary remains disabled

                        checkTertiary = false;

                    }

                }

                continue;

            }

            // Compare primary differences first.

            if ( pSOrder != pTOrder )

            {

                if (sOrder == 0) {

                    // The entire source element is ignorable.

                    // Skip to the next source element, but don't fetch another target element.

                    gett = false;

                    continue;

                }

                if (tOrder == 0) {

                    gets = false;

                    continue;

                }

                // The source and target elements aren't ignorable, but it's still possible

                // for the primary component of one of the elements to be ignorable....

                if (pSOrder == 0)  // primary order in source is ignorable

                {

                    // The source's primary is ignorable, but the target's isn't.  We treat ignorables

                    // as a secondary difference, so remember that we found one.

                    if (checkSecTer) {

                        result = Collator.GREATER;  // (strength is SECONDARY)

                        checkSecTer = false;

                    }

                    // Skip to the next source element, but don't fetch another target element.

                    gett = false;

                }

                else if (pTOrder == 0)

                {

                    // record differences - see the comment above.

                    if (checkSecTer) {

                        result = Collator.LESS;  // (strength is SECONDARY)

                        checkSecTer = false;

                    }

                    // Skip to the next source element, but don't fetch another target element.

                    gets = false;

                } else {

                    // Neither of the orders is ignorable, and we already know that the primary

                    // orders are different because of the (pSOrder != pTOrder) test above.

                    // Record the difference and stop the comparison.

                    if (pSOrder < pTOrder) {

                        return Collator.LESS;  // (strength is PRIMARY)

                    } else {

                        return Collator.GREATER;  // (strength is PRIMARY)

                    }

                }

            } else { // else of if ( pSOrder != pTOrder )

                // primary order is the same, but complete order is different. So there

                // are no base elements at this point, only ignorables (Since the strings are

                // normalized)

                if (checkSecTer) {

                    // a secondary or tertiary difference may still matter

                    short secSOrder = CollationElementIterator.secondaryOrder(sOrder);

                    short secTOrder = CollationElementIterator.secondaryOrder(tOrder);

                    if (secSOrder != secTOrder) {

                        // there is a secondary difference

                        result = (secSOrder < secTOrder) ? Collator.LESS : Collator.GREATER;

                                                // (strength is SECONDARY)

                        checkSecTer = false;

                        // (even in french, only the first secondary difference within

                        //  a base character matters)

                    } else {

                        if (checkTertiary) {

                            // a tertiary difference may still matter

                            short terSOrder = CollationElementIterator.tertiaryOrder(sOrder);

                            short terTOrder = CollationElementIterator.tertiaryOrder(tOrder);

                            if (terSOrder != terTOrder) {

                                // there is a tertiary difference

                                result = (terSOrder < terTOrder) ? Collator.LESS : Collator.GREATER;

                                                // (strength is TERTIARY)

                                checkTertiary = false;

                            }

                        }

                    }

                } // if (checkSecTer)

            }  // if ( pSOrder != pTOrder )

        } // while()

        if (sOrder != CollationElementIterator.NULLORDER) {

            // (tOrder must be CollationElementIterator::NULLORDER,

            //  since this point is only reached when sOrder or tOrder is NULLORDER.)

            // The source string has more elements, but the target string hasn't.

            do {

                if (CollationElementIterator.primaryOrder(sOrder) != 0) {

                    // We found an additional non-ignorable base character in the source string.

                    // This is a primary difference, so the source is greater

                    return Collator.GREATER; // (strength is PRIMARY)

                }