RBCollationTables


1   /*
2    * @(#)RBCollationTables.java   1.9 03/12/19
3    *
4    * Copyright 2004 Sun Microsystems, Inc. All rights reserved.
5    * SUN PROPRIETARY/CONFIDENTIAL. Use is subject to license terms.
6    */
7   
8   /*
9    * (C) Copyright Taligent, Inc. 1996, 1997 - All Rights Reserved
10   * (C) Copyright IBM Corp. 1996-1998 - All Rights Reserved
11   *
12   *   The original version of this source code and documentation is copyrighted
13   * and owned by Taligent, Inc., a wholly-owned subsidiary of IBM. These
14   * materials are provided under terms of a License Agreement between Taligent
15   * and Sun. This technology is protected by multiple US and International
16   * patents. This notice and attribution to Taligent may not be removed.
17   *   Taligent is a registered trademark of Taligent, Inc.
18   *
19   */
20  
21  package java.text;
22  
23  import java.util.Vector  ;
24  import sun.text.UCompactIntArray;
25  import sun.text.IntHashtable;
26  
27  /**
28   * This class contains the static state of a RuleBasedCollator: The various
29   * tables that are used by the collation routines.  Several RuleBasedCollators
30   * can share a single RBCollationTables object, easing memory requirements and
31   * improving performance.
32   */
33  final class RBCollationTables {
34      //===========================================================================================
35      //  The following diagram shows the data structure of the RBCollationTables object.
36      //  Suppose we have the rule, where 'o-umlaut' is the unicode char 0x00F6.
37      //  "a, A < b, B < c, C, ch, cH, Ch, CH < d, D ... < o, O; 'o-umlaut'/E, 'O-umlaut'/E ...".
38      //  What the rule says is, sorts 'ch'ligatures and 'c' only with tertiary difference and
39      //  sorts 'o-umlaut' as if it's always expanded with 'e'.
40      //
41      // mapping table                     contracting list           expanding list
42      // (contains all unicode char
43      //  entries)                   ___    ____________       _________________________
44      //  ________                +>|_*_|->|'c' |v('c') |  +>|v('o')|v('umlaut')|v('e')|
45      // |_\u0001_|-> v('\u0001') | |_:_|  |------------|  | |-------------------------|
46      // |_\u0002_|-> v('\u0002') | |_:_|  |'ch'|v('ch')|  | |             :           |
47      // |____:___|               | |_:_|  |------------|  | |-------------------------|
48      // |____:___|               |        |'cH'|v('cH')|  | |             :           |
49      // |__'a'___|-> v('a')      |        |------------|  | |-------------------------|
50      // |__'b'___|-> v('b')      |        |'Ch'|v('Ch')|  | |             :           |
51      // |____:___|               |        |------------|  | |-------------------------|
52      // |____:___|               |        |'CH'|v('CH')|  | |             :           |
53      // |___'c'__|----------------         ------------   | |-------------------------|
54      // |____:___|                                        | |             :           |
55      // |o-umlaut|----------------------------------------  |_________________________|
56      // |____:___|
57      //
58      // Noted by Helena Shih on 6/23/97
59      //============================================================================================
60  
61      public RBCollationTables(String   rules, int decmp) throws ParseException   {
62          this.rules = rules;
63  
64          RBTableBuilder   builder = new RBTableBuilder  (new BuildAPI());
65          builder.build(rules, decmp); // this object is filled in through
66                                              // the BuildAPI object
67      }
68  
69      final class BuildAPI {
70          /**
71           * Private constructor.  Prevents anyone else besides RBTableBuilder
72           * from gaining direct access to the internals of this class.
73           */
74          private BuildAPI() {
75          }
76  
77          /**
78           * This function is used by RBTableBuilder to fill in all the members of this
79           * object.  (Effectively, the builder class functions as a "friend" of this
80           * class, but to avoid changing too much of the logic, it carries around "shadow"
81           * copies of all these variables until the end of the build process and then
82           * copies them en masse into the actual tables object once all the construction
83           * logic is complete.  This function does that "copying en masse".
84           * @param f2ary The value for frenchSec (the French-secondary flag)
85       * @param swap The value for SE Asian swapping rule
86           * @param map The collator's character-mapping table (the value for mapping)
87           * @param cTbl The collator's contracting-character table (the value for contractTable)
88           * @param eTbl The collator's expanding-character table (the value for expandTable)
89           * @param cFlgs The hash table of characters that participate in contracting-
90           *              character sequences (the value for contractFlags)
91           * @param mso The value for maxSecOrder
92           * @param mto The value for maxTerOrder
93           */
94          void fillInTables(boolean f2ary,
95                boolean swap,
96                            UCompactIntArray map,
97                            Vector   cTbl,
98                            Vector   eTbl,
99                            IntHashtable cFlgs,
100                           short mso,
101                           short mto) {
102             frenchSec = f2ary;
103         seAsianSwapping = swap;
104             mapping = map;
105             contractTable = cTbl;
106             expandTable = eTbl;
107             contractFlags = cFlgs;
108             maxSecOrder = mso;
109             maxTerOrder = mto;
110         }
111     }
112 
113     /**
114      * Gets the table-based rules for the collation object.
115      * @return returns the collation rules that the table collation object
116      * was created from.
117      */
118     public String   getRules()
119     {
120         return rules;
121     }
122 
123     public boolean isFrenchSec() {
124         return frenchSec;
125     }
126     
127     public boolean isSEAsianSwapping() {
128         return seAsianSwapping;
129     }
130     
131     // ==============================================================      
132     // internal (for use by CollationElementIterator)
133     // ==============================================================
134 
135     /**
136      *  Get the entry of hash table of the contracting string in the collation
137      *  table.
138      *  @param ch the starting character of the contracting string
139      */
140     Vector   getContractValues(int ch)
141     {
142         int index = mapping.elementAt(ch);
143         return getContractValuesImpl(index - CONTRACTCHARINDEX);
144     }
145 
146     //get contract values from contractTable by index
147     private Vector   getContractValuesImpl(int index)
148     {
149         if (index >= 0)
150         {
151             return (Vector  )contractTable.elementAt(index);
152         }
153         else // not found
154         {
155             return null;
156         }
157     }
158 
159     /**
160      * Returns true if this character appears anywhere in a contracting
161      * character sequence.  (Used by CollationElementIterator.setOffset().)
162      */
163     boolean usedInContractSeq(int c) {
164         return contractFlags.get(c) == 1;
165     }
166 
167     /**
168       * Return the maximum length of any expansion sequences that end
169       * with the specified comparison order.
170       *
171       * @param order a collation order returned by previous or next.
172       * @return the maximum length of any expansion seuences ending
173       *         with the specified order.
174       *
175       * @see CollationElementIterator#getMaxExpansion
176       */
177     int getMaxExpansion(int order)
178     {
179         int result = 1;
180 
181         if (expandTable != null) {
182             // Right now this does a linear search through the entire
183             // expandsion table.  If a collator had a large number of expansions,
184             // this could cause a performance problem, but in practise that
185             // rarely happens
186             for (int i = 0; i < expandTable.size(); i++) {
187                 int[] valueList = (int [])expandTable.elementAt(i);
188                 int length = valueList.length;
189 
190                 if (length > result && valueList[length-1] == order) {
191                     result = length;
192                 }
193             }
194         }
195 
196         return result;
197     }
198 
199     /**
200      *  Get the entry of hash table of the expanding string in the collation
201      *  table.
202      *  @param idx the index of the expanding string value list
203      */
204     final int[] getExpandValueList(int order) {
205         return (int[])expandTable.elementAt(order - EXPANDCHARINDEX);
206     }
207 
208     /**
209      *  Get the comarison order of a character from the collation table.
210      *  @return the comparison order of a character.
211      */
212     int getUnicodeOrder(int ch)
213     {
214         return mapping.elementAt(ch);
215     }
216 
217     short getMaxSecOrder() {
218         return maxSecOrder;
219     }
220 
221     short getMaxTerOrder() {
222         return maxTerOrder;
223     }
224 
225     /**
226      * Reverse a string.
227      */
228     //shemran/Note: this is used for secondary order value reverse, no
229     //              need to consider supplementary pair.
230     static void reverse (StringBuffer   result, int from, int to)
231     {
232         int i = from;
233         char swap;
234 
235         int j = to - 1;
236         while (i < j) {
237             swap =  result.charAt(i);
238             result.setCharAt(i, result.charAt(j));
239             result.setCharAt(j, swap);
240             i++;
241             j--;
242         }
243     }
244 
245     final static int getEntry(Vector   list, String   name, boolean fwd) {
246         for (int i = 0; i < list.size(); i++) {
247             EntryPair   pair = (EntryPair  )list.elementAt(i);
248             if (pair.fwd == fwd && pair.entryName.equals(name)) {
249                 return i;
250             }
251         }
252         return UNMAPPED;
253     }
254 
255     // ==============================================================
256     // constants
257     // ==============================================================
258     //sherman/Todo: is the value big enough?????
259     final static int EXPANDCHARINDEX = 0x7E000000; // Expand index follows
260     final static int CONTRACTCHARINDEX = 0x7F000000;  // contract indexes follow
261     final static int UNMAPPED = 0xFFFFFFFF;
262 
263     final static int PRIMARYORDERMASK = 0xffff0000;
264     final static int SECONDARYORDERMASK = 0x0000ff00;
265     final static int TERTIARYORDERMASK = 0x000000ff;
266     final static int PRIMARYDIFFERENCEONLY = 0xffff0000;
267     final static int SECONDARYDIFFERENCEONLY = 0xffffff00;
268     final static int PRIMARYORDERSHIFT = 16;
269     final static int SECONDARYORDERSHIFT = 8;
270 
271     // ==============================================================
272     // instance variables
273     // ==============================================================
274     private String   rules = null;
275     private boolean frenchSec = false;
276     private boolean seAsianSwapping = false;
277 
278     private UCompactIntArray mapping = null;
279     private Vector   contractTable = null;
280     private Vector   expandTable = null;
281     private IntHashtable contractFlags = null;
282 
283     private short maxSecOrder = 0;
284     private short maxTerOrder = 0;
285 }
286
A to Z: JavaDoc & Examples Daily Java News & Articles Open Source Projects Open Source Codes Free Computer Books Remove Frame
Popular Tags