_source/plugins/entities/plugin.js

   1 /*
   2 Copyright (c) 2003-2011, CKSource - Frederico Knabben. All rights reserved.
   3 For licensing, see LICENSE.html or http://ckeditor.com/license
   4 */
   5
   6 (function()
   7 {
   8         // Base HTML entities.
   9         var htmlbase = 'nbsp,gt,lt,amp';
  10
  11         var entities =
  12                 // Latin-1 Entities
  13                 'quot,iexcl,cent,pound,curren,yen,brvbar,sect,uml,copy,ordf,laquo,' +
  14                 'not,shy,reg,macr,deg,plusmn,sup2,sup3,acute,micro,para,middot,' +
  15                 'cedil,sup1,ordm,raquo,frac14,frac12,frac34,iquest,times,divide,' +
  16
  17                 // Symbols
  18                 'fnof,bull,hellip,prime,Prime,oline,frasl,weierp,image,real,trade,' +
  19                 'alefsym,larr,uarr,rarr,darr,harr,crarr,lArr,uArr,rArr,dArr,hArr,' +
  20                 'forall,part,exist,empty,nabla,isin,notin,ni,prod,sum,minus,lowast,' +
  21                 'radic,prop,infin,ang,and,or,cap,cup,int,there4,sim,cong,asymp,ne,' +
  22                 'equiv,le,ge,sub,sup,nsub,sube,supe,oplus,otimes,perp,sdot,lceil,' +
  23                 'rceil,lfloor,rfloor,lang,rang,loz,spades,clubs,hearts,diams,' +
  24
  25                 // Other Special Characters
  26                 'circ,tilde,ensp,emsp,thinsp,zwnj,zwj,lrm,rlm,ndash,mdash,lsquo,' +
  27                 'rsquo,sbquo,ldquo,rdquo,bdquo,dagger,Dagger,permil,lsaquo,rsaquo,' +
  28                 'euro';
  29
  30         // Latin Letters Entities
  31         var latin =
  32                 'Agrave,Aacute,Acirc,Atilde,Auml,Aring,AElig,Ccedil,Egrave,Eacute,' +
  33                 'Ecirc,Euml,Igrave,Iacute,Icirc,Iuml,ETH,Ntilde,Ograve,Oacute,Ocirc,' +
  34                 'Otilde,Ouml,Oslash,Ugrave,Uacute,Ucirc,Uuml,Yacute,THORN,szlig,' +
  35                 'agrave,aacute,acirc,atilde,auml,aring,aelig,ccedil,egrave,eacute,' +
  36                 'ecirc,euml,igrave,iacute,icirc,iuml,eth,ntilde,ograve,oacute,ocirc,' +
  37                 'otilde,ouml,oslash,ugrave,uacute,ucirc,uuml,yacute,thorn,yuml,' +
  38                 'OElig,oelig,Scaron,scaron,Yuml';
  39
  40         // Greek Letters Entities.
  41         var greek =
  42                 'Alpha,Beta,Gamma,Delta,Epsilon,Zeta,Eta,Theta,Iota,Kappa,Lambda,Mu,' +
  43                 'Nu,Xi,Omicron,Pi,Rho,Sigma,Tau,Upsilon,Phi,Chi,Psi,Omega,alpha,' +
  44                 'beta,gamma,delta,epsilon,zeta,eta,theta,iota,kappa,lambda,mu,nu,xi,' +
  45                 'omicron,pi,rho,sigmaf,sigma,tau,upsilon,phi,chi,psi,omega,thetasym,' +
  46                 'upsih,piv';
  47
  48         /**
  49          * Create a mapping table between one character and it's entity form from a list of entity names.
  50          * @param reverse {Boolean} Whether create a reverse map from the entity string form to actual character.
  51          */
  52         function buildTable( entities, reverse )
  53         {
  54                 var table = {},
  55                         regex = [];
  56
  57                 // Entities that the browsers DOM don't transform to the final char
  58                 // automatically.
  59                 var specialTable =
  60                         {
  61                                 nbsp    : '\u00A0',             // IE | FF
  62                                 shy             : '\u00AD',             // IE
  63                                 gt              : '\u003E',             // IE | FF |   --   | Opera
  64                                 lt              : '\u003C',             // IE | FF | Safari | Opera
  65                                 amp : '\u0026'          // ALL
  66                         };
  67
  68                 entities = entities.replace( /\b(nbsp|shy|gt|lt|amp)(?:,|$)/g, function( match, entity )
  69                         {
  70                                 var org = reverse ? '&' + entity + ';' : specialTable[ entity ],
  71                                         result = reverse ? specialTable[ entity ] : '&' + entity + ';';
  72
  73                                 table[ org ] = result;
  74                                 regex.push( org );
  75                                 return '';
  76                         });
  77
  78                 if ( !reverse && entities )
  79                 {
  80                         // Transforms the entities string into an array.
  81                         entities = entities.split( ',' );
  82
  83                         // Put all entities inside a DOM element, transforming them to their
  84                         // final chars.
  85                         var div = document.createElement( 'div' ),
  86                                 chars;
  87                         div.innerHTML = '&' + entities.join( ';&' ) + ';';
  88                         chars = div.innerHTML;
  89                         div = null;
  90
  91                         // Add all chars to the table.
  92                         for ( var i = 0 ; i < chars.length ; i++ )
  93                         {
  94                                 var charAt = chars.charAt( i );
  95                                 table[ charAt ] = '&' + entities[ i ] + ';';
  96                                 regex.push( charAt );
  97                         }
  98                 }
  99
 100                 table.regex = regex.join( reverse ? '|' : '' );
 101
 102                 return table;
 103         }
 104
 105         CKEDITOR.plugins.add( 'entities',
 106         {
 107                 afterInit : function( editor )
 108                 {
 109                         var config = editor.config;
 110
 111                         var dataProcessor = editor.dataProcessor,
 112                                 htmlFilter = dataProcessor && dataProcessor.htmlFilter;
 113
 114                         if ( htmlFilter )
 115                         {
 116                                 // Mandatory HTML base entities.
 117                                 var selectedEntities = '';
 118
 119                                 if ( config.basicEntities !== false )
 120                                         selectedEntities += htmlbase;
 121
 122                                 if ( config.entities )
 123                                 {
 124                                         selectedEntities += ',' + entities;
 125                                         if ( config.entities_latin )
 126                                                 selectedEntities += ',' + latin;
 127
 128                                         if ( config.entities_greek )
 129                                                 selectedEntities += ',' + greek;
 130
 131                                         if ( config.entities_additional )
 132                                                 selectedEntities += ',' + config.entities_additional;
 133                                 }
 134
 135                                 var entitiesTable = buildTable( selectedEntities );
 136
 137                                 // Create the Regex used to find entities in the text, leave it matches nothing if entities are empty.
 138                                 var entitiesRegex = entitiesTable.regex ? '[' + entitiesTable.regex + ']' : 'a^';
 139                                 delete entitiesTable.regex;
 140
 141                                 if ( config.entities && config.entities_processNumerical )
 142                                         entitiesRegex = '[^ -~]|' + entitiesRegex ;
 143
 144                                 entitiesRegex = new RegExp( entitiesRegex, 'g' );
 145
 146                                 function getEntity( character )
 147                                 {
 148                                         return config.entities_processNumerical == 'force' || !entitiesTable[ character ] ?
 149                                                    '&#' + character.charCodeAt(0) + ';'
 150                                                         : entitiesTable[ character ];
 151                                 }
 152
 153                                 // Decode entities that the browsers has transformed
 154                                 // at first place.
 155                                 var baseEntitiesTable = buildTable( [ htmlbase, 'shy' ].join( ',' ) , true ),
 156                                         baseEntitiesRegex = new RegExp( baseEntitiesTable.regex, 'g' );
 157
 158                                 function getChar( character )
 159                                 {
 160                                         return baseEntitiesTable[ character ];
 161                                 }
 162
 163                                 htmlFilter.addRules(
 164                                         {
 165                                                 text : function( text )
 166                                                 {
 167                                                         return text.replace( baseEntitiesRegex, getChar )
 168                                                                         .replace( entitiesRegex, getEntity );
 169                                                 }
 170                                         });
 171                         }
 172                 }
 173         });
 174 })();
 175
 176 /**
 177  * Whether to escape HTML preserved entities in text, including:
 178  * <ul>
 179  * <li>nbsp</li>
 180  * <li>gt</li>
 181  * <li>lt</li>
 182  * <li>amp</li>
 183  * </ul>
 184  * <strong>Note:</strong> It should not be subjected to change unless you're outputting non-HTML data format like BBCode.
 185  * @type Boolean
 186  * @default true
 187  * @example
 188  * config.basicEntities = false;
 189  */
 190 CKEDITOR.config.basicEntities = true;
 191
 192 /**
 193  * Whether to use HTML entities in the output.
 194  * @name CKEDITOR.config.entities
 195  * @type Boolean
 196  * @default true
 197  * @example
 198  * config.entities = false;
 199  */
 200 CKEDITOR.config.entities = true;
 201
 202 /**
 203  * Whether to convert some Latin characters (Latin alphabet No&#46; 1, ISO 8859-1)
 204  * to HTML entities. The list of entities can be found at the
 205  * <a href="http://www.w3.org/TR/html4/sgml/entities.html#h-24.2.1">W3C HTML 4.01 Specification, section 24.2.1</a>.
 206  * @name CKEDITOR.config.entities_latin
 207  * @type Boolean
 208  * @default true
 209  * @example
 210  * config.entities_latin = false;
 211  */
 212 CKEDITOR.config.entities_latin = true;
 213
 214 /**
 215  * Whether to convert some symbols, mathematical symbols, and Greek letters to
 216  * HTML entities. This may be more relevant for users typing text written in Greek.
 217  * The list of entities can be found at the
 218  * <a href="http://www.w3.org/TR/html4/sgml/entities.html#h-24.3.1">W3C HTML 4.01 Specification, section 24.3.1</a>.
 219  * @name CKEDITOR.config.entities_greek
 220  * @type Boolean
 221  * @default true
 222  * @example
 223  * config.entities_greek = false;
 224  */
 225 CKEDITOR.config.entities_greek = true;
 226
 227 /**
 228  * Whether to convert all remaining characters, not comprised in the ASCII
 229  * character table, to their relative decimal numeric representation of HTML entity.
 230  * When specified as the value 'force', it will simply convert all entities into the above form.
 231  * For example, the phrase "This is Chinese: &#27721;&#35821;." is outputted
 232  * as "This is Chinese: &amp;#27721;&amp;#35821;."
 233  * @name CKEDITOR.config.entities_processNumerical
 234  * @type Boolean|String
 235  * @default false
 236  * @example
 237  * config.entities_processNumerical = true;
 238  * config.entities_processNumerical = 'force';          //Convert from "&nbsp;" into "&#160;";
 239  */
 240
 241 /**
 242  * An additional list of entities to be used. It's a string containing each
 243  * entry separated by a comma. Entities names or number must be used, exclusing
 244  * the "&amp;" preffix and the ";" termination.
 245  * @name CKEDITOR.config.entities_additional
 246  * @default '#39'  // The single quote (') character.
 247  * @type String
 248  * @example
 249  */
 250 CKEDITOR.config.entities_additional = '#39';