| 1 | /* |
| 2 | * jDTAUS Core Utilities |
| 3 | * Copyright (C) 2012 Christian Schulte |
| 4 | * <cs@schulte.it> |
| 5 | * |
| 6 | * This library is free software; you can redistribute it and/or |
| 7 | * modify it under the terms of the GNU Lesser General Public |
| 8 | * License as published by the Free Software Foundation; either |
| 9 | * version 2.1 of the License, or any later version. |
| 10 | * |
| 11 | * This library is distributed in the hope that it will be useful, |
| 12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 14 | * Lesser General Public License for more details. |
| 15 | * |
| 16 | * You should have received a copy of the GNU Lesser General Public |
| 17 | * License along with this library; if not, write to the Free Software |
| 18 | * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA |
| 19 | * |
| 20 | */ |
| 21 | package org.jdtaus.core.text.util; |
| 22 | |
| 23 | import java.io.IOException; |
| 24 | import java.io.InputStream; |
| 25 | import java.lang.ref.Reference; |
| 26 | import java.lang.ref.SoftReference; |
| 27 | import java.net.URL; |
| 28 | import java.util.Collections; |
| 29 | import java.util.Enumeration; |
| 30 | import java.util.HashMap; |
| 31 | import java.util.Map; |
| 32 | import java.util.Properties; |
| 33 | |
| 34 | /** |
| 35 | * Provides static methods for encoding/decoding HTML. |
| 36 | * |
| 37 | * @author <a href="mailto:cs@schulte.it">Christian Schulte</a> |
| 38 | * @version $JDTAUS: HtmlEntities.java 8641 2012-09-27 06:45:17Z schulte $ |
| 39 | * @since 1.11 |
| 40 | */ |
| 41 | class HtmlEntities |
| 42 | { |
| 43 | |
| 44 | /** Mapping of HTML entity names to characters. */ |
| 45 | private static volatile Reference/*<Map<String,Character>>*/ entityMap; |
| 46 | |
| 47 | /** Mapping of characters to HTML entity names. */ |
| 48 | private static volatile Reference/*<Map<Character,String>>*/ characterMap; |
| 49 | |
| 50 | /** Constant for the prefix of keys of HTML entity properties. */ |
| 51 | private static final String PROPERTY_PREFIX = "HtmlEntities.entityName."; |
| 52 | |
| 53 | /** |
| 54 | * Decodes an entity name to a character. |
| 55 | * |
| 56 | * @param entityName The name of the entity to decode. |
| 57 | * |
| 58 | * @return The character represented by {@code entityName} or {@code null}, if no such character is found. |
| 59 | * |
| 60 | * @throws NullPointerException if {@code entityName} is {@code null}. |
| 61 | */ |
| 62 | static Character toCharacter( final String entityName ) |
| 63 | { |
| 64 | if ( entityName == null ) |
| 65 | { |
| 66 | throw new NullPointerException( "entityName" ); |
| 67 | } |
| 68 | |
| 69 | Map map = entityMap != null ? (Map) entityMap.get() : null; |
| 70 | |
| 71 | if ( map == null ) |
| 72 | { |
| 73 | map = new HashMap/*<String,Character>*/( 512 ); |
| 74 | InputStream in = null; |
| 75 | boolean close = true; |
| 76 | |
| 77 | final URL resources = HtmlEntities.class.getResource( "HtmlEntities.properties" ); |
| 78 | final Properties properties = new Properties(); |
| 79 | |
| 80 | try |
| 81 | { |
| 82 | in = resources.openStream(); |
| 83 | properties.load( in ); |
| 84 | in.close(); |
| 85 | close = false; |
| 86 | } |
| 87 | catch ( final IOException e ) |
| 88 | { |
| 89 | throw new AssertionError( e ); |
| 90 | } |
| 91 | finally |
| 92 | { |
| 93 | try |
| 94 | { |
| 95 | if ( close && in != null ) |
| 96 | { |
| 97 | in.close(); |
| 98 | } |
| 99 | } |
| 100 | catch ( final IOException e ) |
| 101 | { |
| 102 | throw new AssertionError( e ); |
| 103 | } |
| 104 | } |
| 105 | |
| 106 | for ( final Enumeration e = properties.propertyNames(); e.hasMoreElements(); ) |
| 107 | { |
| 108 | final String name = (String) e.nextElement(); |
| 109 | final String value = properties.getProperty( name ); |
| 110 | |
| 111 | if ( name.startsWith( PROPERTY_PREFIX ) ) |
| 112 | { |
| 113 | map.put( name.substring( PROPERTY_PREFIX.length() ), new Character( value.charAt( 0 ) ) ); |
| 114 | } |
| 115 | } |
| 116 | |
| 117 | entityMap = new SoftReference( Collections.synchronizedMap( map ) ); |
| 118 | } |
| 119 | |
| 120 | return (Character) map.get( entityName ); |
| 121 | } |
| 122 | |
| 123 | /** |
| 124 | * Encodes a character to an entity name. |
| 125 | * |
| 126 | * @param character The character to encode. |
| 127 | * |
| 128 | * @return The entity name representing {@code character} or {@code null}, if no such entity name is found. |
| 129 | * |
| 130 | * @throws NullPointerException if {@code character} is {@code null}. |
| 131 | */ |
| 132 | static String toEntity( final Character character ) |
| 133 | { |
| 134 | if ( character == null ) |
| 135 | { |
| 136 | throw new NullPointerException( "character" ); |
| 137 | } |
| 138 | |
| 139 | Map map = characterMap != null ? (Map) characterMap.get() : null; |
| 140 | |
| 141 | if ( map == null ) |
| 142 | { |
| 143 | map = new HashMap/*<Character,String>*/( 512 ); |
| 144 | InputStream in = null; |
| 145 | boolean close = true; |
| 146 | |
| 147 | final URL resources = HtmlEntities.class.getResource( "HtmlEntities.properties" ); |
| 148 | final Properties properties = new Properties(); |
| 149 | |
| 150 | try |
| 151 | { |
| 152 | in = resources.openStream(); |
| 153 | properties.load( in ); |
| 154 | in.close(); |
| 155 | close = false; |
| 156 | } |
| 157 | catch ( final IOException e ) |
| 158 | { |
| 159 | throw new AssertionError( e ); |
| 160 | } |
| 161 | finally |
| 162 | { |
| 163 | try |
| 164 | { |
| 165 | if ( close && in != null ) |
| 166 | { |
| 167 | in.close(); |
| 168 | } |
| 169 | } |
| 170 | catch ( final IOException e ) |
| 171 | { |
| 172 | throw new AssertionError( e ); |
| 173 | } |
| 174 | } |
| 175 | |
| 176 | for ( final Enumeration e = properties.propertyNames(); e.hasMoreElements(); ) |
| 177 | { |
| 178 | final String name = (String) e.nextElement(); |
| 179 | final String value = properties.getProperty( name ); |
| 180 | |
| 181 | if ( name.startsWith( PROPERTY_PREFIX ) ) |
| 182 | { |
| 183 | map.put( new Character( value.charAt( 0 ) ), name.substring( PROPERTY_PREFIX.length() ) ); |
| 184 | } |
| 185 | } |
| 186 | |
| 187 | characterMap = new SoftReference( Collections.synchronizedMap( map ) ); |
| 188 | } |
| 189 | |
| 190 | return (String) map.get( character ); |
| 191 | } |
| 192 | |
| 193 | /** |
| 194 | * Encodes a string to HTML. |
| 195 | * |
| 196 | * @param str The string to encode or {@code null}. |
| 197 | * |
| 198 | * @return {@code str} encoded to HTML or {@code null}. |
| 199 | */ |
| 200 | static String escapeHtml( final String str ) |
| 201 | { |
| 202 | String encoded = null; |
| 203 | |
| 204 | if ( str != null ) |
| 205 | { |
| 206 | final StringBuffer b = new StringBuffer( str.length() ); |
| 207 | |
| 208 | for ( int i = 0, s0 = str.length(); i < s0; i++ ) |
| 209 | { |
| 210 | final Character c = new Character( str.charAt( i ) ); |
| 211 | final String entityName = toEntity( c ); |
| 212 | |
| 213 | if ( entityName != null ) |
| 214 | { |
| 215 | b.append( '&' ).append( entityName ).append( ';' ); |
| 216 | } |
| 217 | else |
| 218 | { |
| 219 | b.append( c ); |
| 220 | } |
| 221 | } |
| 222 | |
| 223 | encoded = b.toString(); |
| 224 | } |
| 225 | |
| 226 | return encoded; |
| 227 | } |
| 228 | |
| 229 | /** |
| 230 | * Decodes HTML to a string. |
| 231 | * |
| 232 | * @param html The HTML to decode or {@code null}. |
| 233 | * |
| 234 | * @return {@code html} decoded or {@code null}. |
| 235 | */ |
| 236 | static String unescapeHtml( final String html ) |
| 237 | { |
| 238 | String decoded = null; |
| 239 | |
| 240 | if ( html != null ) |
| 241 | { |
| 242 | final StringBuffer b = new StringBuffer( html.length() ); |
| 243 | final StringBuffer entityName = new StringBuffer( 8 ); |
| 244 | boolean parsingEntityName = false; |
| 245 | |
| 246 | for ( int i = 0, s0 = html.length(); i < s0; i++ ) |
| 247 | { |
| 248 | final char c = html.charAt( i ); |
| 249 | |
| 250 | if ( c == '&' ) |
| 251 | { |
| 252 | parsingEntityName = true; |
| 253 | entityName.setLength( 0 ); |
| 254 | } |
| 255 | |
| 256 | if ( parsingEntityName ) |
| 257 | { |
| 258 | if ( c == ';' ) |
| 259 | { |
| 260 | final Character character = toCharacter( entityName.toString() ); |
| 261 | |
| 262 | if ( character != null ) |
| 263 | { |
| 264 | b.append( character ); |
| 265 | } |
| 266 | else |
| 267 | { |
| 268 | b.append( '&' ).append( entityName.toString() ).append( ';' ); |
| 269 | } |
| 270 | |
| 271 | parsingEntityName = false; |
| 272 | } |
| 273 | else |
| 274 | { |
| 275 | entityName.append( c ); |
| 276 | } |
| 277 | } |
| 278 | else |
| 279 | { |
| 280 | b.append( c ); |
| 281 | } |
| 282 | } |
| 283 | |
| 284 | decoded = b.toString(); |
| 285 | } |
| 286 | |
| 287 | return decoded; |
| 288 | } |
| 289 | |
| 290 | } |