platform/lang-impl/src/com/intellij/psi/impl/source/CharTableImpl.java

   1 /*
   2  * Copyright 2000-2009 JetBrains s.r.o.
   3  *
   4  * Licensed under the Apache License, Version 2.0 (the "License");
   5  * you may not use this file except in compliance with the License.
   6  * You may obtain a copy of the License at
   7  *
   8  * http://www.apache.org/licenses/LICENSE-2.0
   9  *
  10  * Unless required by applicable law or agreed to in writing, software
  11  * distributed under the License is distributed on an "AS IS" BASIS,
  12  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  13  * See the License for the specific language governing permissions and
  14  * limitations under the License.
  15  */
  16
  17 package com.intellij.psi.impl.source;
  18
  19 import com.intellij.util.CharTable;
  20 import com.intellij.util.text.CharArrayCharSequence;
  21 import com.intellij.util.text.CharArrayUtil;
  22 import com.intellij.util.text.CharSequenceHashingStrategy;
  23 import gnu.trove.THashSet;
  24
  25 /**
  26  * @author max
  27  */
  28 public class CharTableImpl implements CharTable {
  29   private static final int INTERN_THRESHOLD = 40; // 40 or more characters long tokens won't be interned.
  30   private static final CharSequenceHashingStrategy HASHER = new CharSequenceHashingStrategy();
  31   private static final MyTHashSet staticEntries = new MyStaticTHashSet();
  32   private final MyTHashSet entries = new MyTHashSet();
  33
  34   public CharSequence intern(final CharSequence text) {
  35     if (text.length() > INTERN_THRESHOLD) return createSequence(text);
  36
  37     int idx = staticEntries.index(text);
  38     if (idx >= 0) {
  39       //noinspection NonPrivateFieldAccessedInSynchronizedContext
  40       return staticEntries.get(idx);
  41     }
  42
  43     synchronized(this) {
  44       idx = entries.index(text);
  45       if (idx >= 0) {
  46         //noinspection NonPrivateFieldAccessedInSynchronizedContext
  47         return entries.get(idx);
  48       }
  49
  50       // We need to create separate string just to prevent referencing all character data when original is string or char sequence over string
  51       final CharSequence entry = createSequence(text);
  52       boolean added = entries.add(entry);
  53       assert added;
  54
  55       return entry;
  56     }
  57   }
  58
  59   public CharSequence intern(CharSequence baseText, int startOffset, int endOffset) {
  60     if (endOffset - startOffset == baseText.length()) return baseText.toString();
  61     return intern(baseText.subSequence(startOffset, endOffset));
  62   }
  63
  64   private static CharSequence createSequence(final CharSequence text) {
  65     final char[] buf = new char[text.length()];
  66     CharArrayUtil.getChars(text, buf, 0);
  67     return new CharArrayCharSequence(buf);
  68   }
  69
  70   public static void staticIntern(final String text) {
  71     synchronized(staticEntries) {
  72       staticEntries.add(text);
  73     }
  74   }
  75
  76   private static class MyTHashSet extends THashSet<CharSequence> {
  77     private MyTHashSet() {
  78       super(10, 0.9f, HASHER);
  79     }
  80
  81     public int index(final CharSequence obj) {
  82       return super.index(obj);
  83     }
  84
  85     public CharSequence get(int index) {
  86       return (CharSequence)_set[index];
  87     }
  88   }
  89
  90   private static class MyStaticTHashSet extends MyTHashSet {{
  91     add("==" );
  92     add("!=" );
  93     add("||" );
  94     add("++" );
  95     add("--" );
  96
  97     add("<" );
  98     add("<=" );
  99     add("<<=" );
 100     add("<<" );
 101     add(">" );
 102     add("&" );
 103     add("&&" );
 104
 105     add("+=" );
 106     add("-=" );
 107     add("*=" );
 108     add("/=" );
 109     add("&=" );
 110     add("|=" );
 111     add("^=" );
 112     add("%=" );
 113
 114     add("("   );
 115     add(")"   );
 116     add("{"   );
 117     add("}"   );
 118     add("["   );
 119     add("]"   );
 120     add(";"   );
 121     add(","   );
 122     add("..." );
 123     add("."   );
 124
 125     add("=" );
 126     add("!" );
 127     add("~" );
 128     add("?" );
 129     add(":" );
 130     add("+" );
 131     add("-" );
 132     add("*" );
 133     add("/" );
 134     add("|" );
 135     add("^" );
 136     add("%" );
 137     add("@" );
 138
 139     add(" " );
 140     add("  " );
 141     add("   " );
 142     add("    " );
 143     add("     " );
 144     add("      " );
 145     add("       " );
 146     add("        " );
 147     add("         " );
 148     add("          " );
 149     add("           " );
 150     add("            " );
 151     add("             " );
 152     add("              " );
 153     add("               " );
 154     add("\n" );
 155     add("\n  " );
 156     add("\n    " );
 157     add("\n      " );
 158     add("\n        " );
 159     add("\n          " );
 160     add("\n            " );
 161     add("\n              " );
 162     add("\n                " );
 163
 164     add("<");
 165     add(">");
 166     add("</");
 167     add("/>");
 168     add("\"");
 169     add("\'");
 170     add("<![CDATA[");
 171     add("]]>");
 172     add("<!--");
 173     add("-->");
 174     add("<!DOCTYPE");
 175     add("SYSTEM");
 176     add("PUBLIC");
 177     add("<?");
 178     add("?>");
 179
 180     add("<%");
 181     add("%>");
 182     add("<%=");
 183     add("<%@");
 184     add("${");
 185     add("");
 186   }}
 187 }