sources/plugins/entities/plugin.js

   1 /**
   2  * @license Copyright (c) 2003-2015, CKSource - Frederico Knabben. All rights reserved.
   3  * For licensing, see LICENSE.md or http://ckeditor.com/license
   4  */
   5
   6 ( function() {
   7         // Basic HTML entities.
   8         var htmlbase = 'nbsp,gt,lt,amp';
   9
  10         var entities =
  11         // Latin-1 entities
  12         'quot,iexcl,cent,pound,curren,yen,brvbar,sect,uml,copy,ordf,laquo,' +
  13                 'not,shy,reg,macr,deg,plusmn,sup2,sup3,acute,micro,para,middot,' +
  14                 'cedil,sup1,ordm,raquo,frac14,frac12,frac34,iquest,times,divide,' +
  15
  16                 // Symbols
  17                 'fnof,bull,hellip,prime,Prime,oline,frasl,weierp,image,real,trade,' +
  18                 'alefsym,larr,uarr,rarr,darr,harr,crarr,lArr,uArr,rArr,dArr,hArr,' +
  19                 'forall,part,exist,empty,nabla,isin,notin,ni,prod,sum,minus,lowast,' +
  20                 'radic,prop,infin,ang,and,or,cap,cup,int,there4,sim,cong,asymp,ne,' +
  21                 'equiv,le,ge,sub,sup,nsub,sube,supe,oplus,otimes,perp,sdot,lceil,' +
  22                 'rceil,lfloor,rfloor,lang,rang,loz,spades,clubs,hearts,diams,' +
  23
  24                 // Other special characters
  25                 'circ,tilde,ensp,emsp,thinsp,zwnj,zwj,lrm,rlm,ndash,mdash,lsquo,' +
  26                 'rsquo,sbquo,ldquo,rdquo,bdquo,dagger,Dagger,permil,lsaquo,rsaquo,' +
  27                 'euro';
  28
  29         // Latin letters entities
  30         var latin = 'Agrave,Aacute,Acirc,Atilde,Auml,Aring,AElig,Ccedil,Egrave,Eacute,' +
  31                 'Ecirc,Euml,Igrave,Iacute,Icirc,Iuml,ETH,Ntilde,Ograve,Oacute,Ocirc,' +
  32                 'Otilde,Ouml,Oslash,Ugrave,Uacute,Ucirc,Uuml,Yacute,THORN,szlig,' +
  33                 'agrave,aacute,acirc,atilde,auml,aring,aelig,ccedil,egrave,eacute,' +
  34                 'ecirc,euml,igrave,iacute,icirc,iuml,eth,ntilde,ograve,oacute,ocirc,' +
  35                 'otilde,ouml,oslash,ugrave,uacute,ucirc,uuml,yacute,thorn,yuml,' +
  36                 'OElig,oelig,Scaron,scaron,Yuml';
  37
  38         // Greek letters entities.
  39         var greek = 'Alpha,Beta,Gamma,Delta,Epsilon,Zeta,Eta,Theta,Iota,Kappa,Lambda,Mu,' +
  40                 'Nu,Xi,Omicron,Pi,Rho,Sigma,Tau,Upsilon,Phi,Chi,Psi,Omega,alpha,' +
  41                 'beta,gamma,delta,epsilon,zeta,eta,theta,iota,kappa,lambda,mu,nu,xi,' +
  42                 'omicron,pi,rho,sigmaf,sigma,tau,upsilon,phi,chi,psi,omega,thetasym,' +
  43                 'upsih,piv';
  44
  45         // Create a mapping table between one character and its entity form from a list of entity names.
  46         // @param reverse {Boolean} Whether to create a reverse map from the entity string form to an actual character.
  47         function buildTable( entities, reverse ) {
  48                 var table = {},
  49                         regex = [];
  50
  51                 // Entities that the browsers' DOM does not automatically transform to the
  52                 // final character.
  53                 var specialTable = {
  54                         nbsp: '\u00A0', // IE | FF
  55                         shy: '\u00AD', // IE
  56                         gt: '\u003E', // IE | FF |   --   | Opera
  57                         lt: '\u003C', // IE | FF | Safari | Opera
  58                         amp: '\u0026', // ALL
  59                         apos: '\u0027', // IE
  60                         quot: '\u0022' // IE
  61                 };
  62
  63                 entities = entities.replace( /\b(nbsp|shy|gt|lt|amp|apos|quot)(?:,|$)/g, function( match, entity ) {
  64                         var org = reverse ? '&' + entity + ';' : specialTable[ entity ],
  65                                 result = reverse ? specialTable[ entity ] : '&' + entity + ';';
  66
  67                         table[ org ] = result;
  68                         regex.push( org );
  69                         return '';
  70                 } );
  71
  72                 if ( !reverse && entities ) {
  73                         // Transforms the entities string into an array.
  74                         entities = entities.split( ',' );
  75
  76                         // Put all entities inside a DOM element, transforming them to their
  77                         // final characters.
  78                         var div = document.createElement( 'div' ),
  79                                 chars;
  80                         div.innerHTML = '&' + entities.join( ';&' ) + ';';
  81                         chars = div.innerHTML;
  82                         div = null;
  83
  84                         // Add all characters to the table.
  85                         for ( var i = 0; i < chars.length; i++ ) {
  86                                 var charAt = chars.charAt( i );
  87                                 table[ charAt ] = '&' + entities[ i ] + ';';
  88                                 regex.push( charAt );
  89                         }
  90                 }
  91
  92                 table.regex = regex.join( reverse ? '|' : '' );
  93
  94                 return table;
  95         }
  96
  97         CKEDITOR.plugins.add( 'entities', {
  98                 afterInit: function( editor ) {
  99                         var config = editor.config;
 100
 101                         function getChar( character ) {
 102                                 return baseEntitiesTable[ character ];
 103                         }
 104
 105                         function getEntity( character ) {
 106                                 return config.entities_processNumerical == 'force' || !entitiesTable[ character ] ? '&#' + character.charCodeAt( 0 ) + ';'
 107                                 : entitiesTable[ character ];
 108                         }
 109
 110                         var dataProcessor = editor.dataProcessor,
 111                                 htmlFilter = dataProcessor && dataProcessor.htmlFilter;
 112
 113                         if ( htmlFilter ) {
 114                                 // Mandatory HTML basic entities.
 115                                 var selectedEntities = [];
 116
 117                                 if ( config.basicEntities !== false )
 118                                         selectedEntities.push( htmlbase );
 119
 120                                 if ( config.entities ) {
 121                                         if ( selectedEntities.length )
 122                                                 selectedEntities.push( entities );
 123
 124                                         if ( config.entities_latin )
 125                                                 selectedEntities.push( latin );
 126
 127                                         if ( config.entities_greek )
 128                                                 selectedEntities.push( greek );
 129
 130                                         if ( config.entities_additional )
 131                                                 selectedEntities.push( config.entities_additional );
 132                                 }
 133
 134                                 var entitiesTable = buildTable( selectedEntities.join( ',' ) );
 135
 136                                 // Create the Regex used to find entities in the text, leave it matches nothing if entities are empty.
 137                                 var entitiesRegex = entitiesTable.regex ? '[' + entitiesTable.regex + ']' : 'a^';
 138                                 delete entitiesTable.regex;
 139
 140                                 if ( config.entities && config.entities_processNumerical )
 141                                         entitiesRegex = '[^ -~]|' + entitiesRegex;
 142
 143                                 entitiesRegex = new RegExp( entitiesRegex, 'g' );
 144
 145                                 // Decode entities that the browsers has transformed
 146                                 // at first place.
 147                                 var baseEntitiesTable = buildTable( [ htmlbase, 'shy' ].join( ',' ), true ),
 148                                         baseEntitiesRegex = new RegExp( baseEntitiesTable.regex, 'g' );
 149
 150                                 htmlFilter.addRules( {
 151                                         text: function( text ) {
 152                                                 return text.replace( baseEntitiesRegex, getChar ).replace( entitiesRegex, getEntity );
 153                                         }
 154                                 }, {
 155                                         applyToAll: true,
 156                                         excludeNestedEditable: true
 157                                 } );
 158                         }
 159                 }
 160         } );
 161 } )();
 162
 163 /**
 164  * Whether to escape basic HTML entities in the document, including:
 165  *
 166  * * `&nbsp;`
 167  * * `&gt;`
 168  * * `&lt;`
 169  * * `&amp;`
 170  *
 171  * **Note:** This option should not be changed unless when outputting a non-HTML data format like BBCode.
 172  *
 173  *              config.basicEntities = false;
 174  *
 175  * @cfg {Boolean} [basicEntities=true]
 176  * @member CKEDITOR.config
 177  */
 178 CKEDITOR.config.basicEntities = true;
 179
 180 /**
 181  * Whether to use HTML entities in the editor output.
 182  *
 183  *              config.entities = false;
 184  *
 185  * @cfg {Boolean} [entities=true]
 186  * @member CKEDITOR.config
 187  */
 188 CKEDITOR.config.entities = true;
 189
 190 /**
 191  * Whether to convert some Latin characters (Latin alphabet No. 1, ISO 8859-1)
 192  * to HTML entities. The list of entities can be found in the
 193  * [W3C HTML 4.01 Specification, section 24.2.1](http://www.w3.org/TR/html4/sgml/entities.html#h-24.2.1).
 194  *
 195  *              config.entities_latin = false;
 196  *
 197  * @cfg {Boolean} [entities_latin=true]
 198  * @member CKEDITOR.config
 199  */
 200 CKEDITOR.config.entities_latin = true;
 201
 202 /**
 203  * Whether to convert some symbols, mathematical symbols, and Greek letters to
 204  * HTML entities. This may be more relevant for users typing text written in Greek.
 205  * The list of entities can be found in the
 206  * [W3C HTML 4.01 Specification, section 24.3.1](http://www.w3.org/TR/html4/sgml/entities.html#h-24.3.1).
 207  *
 208  *              config.entities_greek = false;
 209  *
 210  * @cfg {Boolean} [entities_greek=true]
 211  * @member CKEDITOR.config
 212  */
 213 CKEDITOR.config.entities_greek = true;
 214
 215 /**
 216  * Whether to convert all remaining characters not included in the ASCII
 217  * character table to their relative decimal numeric representation of HTML entity.
 218  * When set to `force`, it will convert all entities into this format.
 219  *
 220  * For example the phrase: `'This is Chinese: 汉语.'` would be output
 221  * as: `'This is Chinese: &#27721;&#35821;.'`
 222  *
 223  *              config.entities_processNumerical = true;
 224  *              config.entities_processNumerical = 'force'; // Converts from '&nbsp;' into '&#160;';
 225  *
 226  * @cfg {Boolean/String} [entities_processNumerical=false]
 227  * @member CKEDITOR.config
 228  */
 229
 230 /**
 231  * A comma-separated list of  additional entities to be used. Entity names
 232  * or numbers must be used in a form that excludes the `'&amp;'` prefix and the `';'` ending.
 233  *
 234  *              config.entities_additional = '#1049'; // Adds Cyrillic capital letter Short I (Й).
 235  *
 236  * @cfg {String} [entities_additional='#39' (The single quote (') character)]
 237  * @member CKEDITOR.config
 238  */
 239 CKEDITOR.config.entities_additional = '#39';