1/*
2 * FCKeditor - The text editor for Internet - http://www.fckeditor.net
3 * Copyright (C) 2003-2007 Frederico Caldeira Knabben
4 *
5 * == BEGIN LICENSE ==
6 *
7 * Licensed under the terms of any of the following licenses at your
8 * choice:
9 *
10 *  - GNU General Public License Version 2 or later (the "GPL")
11 *    http://www.gnu.org/licenses/gpl.html
12 *
13 *  - GNU Lesser General Public License Version 2.1 or later (the "LGPL")
14 *    http://www.gnu.org/licenses/lgpl.html
15 *
16 *  - Mozilla Public License Version 1.1 or later (the "MPL")
17 *    http://www.mozilla.org/MPL/MPL-1.1.html
18 *
19 * == END LICENSE ==
20 *
21 * Defines the FCKXHtml object, responsible for the XHTML operations.
22 */
23
24var FCKXHtml = new Object() ;
25
26FCKXHtml.CurrentJobNum = 0 ;
27
28FCKXHtml.GetXHTML = function( node, includeNode, format )
29{
30	FCKXHtmlEntities.Initialize() ;
31
32	// Set the correct entity to use for empty blocks.
33	this._NbspEntity = ( FCKConfig.ProcessHTMLEntities? 'nbsp' : '#160' ) ;
34
35	// Save the current IsDirty state. The XHTML processor may change the
36	// original HTML, dirtying it.
37	var bIsDirty = FCK.IsDirty() ;
38
39	this._CreateNode = FCKConfig.ForceStrongEm ? FCKXHtml_CreateNode_StrongEm : FCKXHtml_CreateNode_Normal ;
40
41	// Special blocks are blocks of content that remain untouched during the
42	// process. It is used for SCRIPTs and STYLEs.
43	FCKXHtml.SpecialBlocks = new Array() ;
44
45	// Create the XML DOMDocument object.
46	this.XML = FCKTools.CreateXmlObject( 'DOMDocument' ) ;
47
48	// Add a root element that holds all child nodes.
49	this.MainNode = this.XML.appendChild( this.XML.createElement( 'xhtml' ) ) ;
50
51	FCKXHtml.CurrentJobNum++ ;
52
53	if ( includeNode )
54		this._AppendNode( this.MainNode, node ) ;
55	else
56		this._AppendChildNodes( this.MainNode, node, false ) ;
57
58	// Get the resulting XHTML as a string.
59	var sXHTML = this._GetMainXmlString() ;
60
61	this.XML = null ;
62
63	// Strip the "XHTML" root node.
64	sXHTML = sXHTML.substr( 7, sXHTML.length - 15 ).Trim() ;
65
66	// Remove the trailing <br> added by Gecko.
67	// REMOVE: Maybe the following is not anymore necessary because a similar
68	// check is made on _AppendNode
69	if ( FCKBrowserInfo.IsGecko )
70		sXHTML = sXHTML.replace( /<br\/>$/, '' ) ;
71
72	// Add a space in the tags with no closing tags, like <br/> -> <br />
73	sXHTML = sXHTML.replace( FCKRegexLib.SpaceNoClose, ' />');
74
75	if ( FCKConfig.ForceSimpleAmpersand )
76		sXHTML = sXHTML.replace( FCKRegexLib.ForceSimpleAmpersand, '&' ) ;
77
78	if ( format )
79		sXHTML = FCKCodeFormatter.Format( sXHTML ) ;
80
81	// Now we put back the SpecialBlocks contents.
82	for ( var i = 0 ; i < FCKXHtml.SpecialBlocks.length ; i++ )
83	{
84		var oRegex = new RegExp( '___FCKsi___' + i ) ;
85		sXHTML = sXHTML.replace( oRegex, FCKXHtml.SpecialBlocks[i] ) ;
86	}
87
88	// Replace entities marker with the ampersand.
89	sXHTML = sXHTML.replace( FCKRegexLib.GeckoEntitiesMarker, '&' ) ;
90
91	// Restore the IsDirty state if it was not dirty.
92	if ( !bIsDirty )
93		FCK.ResetIsDirty() ;
94
95	return sXHTML ;
96}
97
98FCKXHtml._AppendAttribute = function( xmlNode, attributeName, attributeValue )
99{
100	try
101	{
102		if ( attributeValue == undefined || attributeValue == null )
103			attributeValue = '' ;
104		else if ( attributeValue.replace )
105		{
106			if ( FCKConfig.ForceSimpleAmpersand )
107				attributeValue = attributeValue.replace( /&/g, '___FCKAmp___' ) ;
108
109			// Entities must be replaced in the attribute values.
110			attributeValue = attributeValue.replace( FCKXHtmlEntities.EntitiesRegex, FCKXHtml_GetEntity ) ;
111		}
112
113		// Create the attribute.
114		var oXmlAtt = this.XML.createAttribute( attributeName ) ;
115		oXmlAtt.value = attributeValue ;
116
117		// Set the attribute in the node.
118		xmlNode.attributes.setNamedItem( oXmlAtt ) ;
119	}
120	catch (e)
121	{}
122}
123
124FCKXHtml._AppendChildNodes = function( xmlNode, htmlNode, isBlockElement )
125{
126	// Trim block elements. This is also needed to avoid Firefox leaving extra
127	// BRs at the end of them.
128	if ( isBlockElement )
129		FCKDomTools.TrimNode( htmlNode, true ) ;
130
131	var iCount = 0 ;
132
133	var oNode = htmlNode.firstChild ;
134
135	while ( oNode )
136	{
137		if ( this._AppendNode( xmlNode, oNode ) )
138			iCount++ ;
139
140		oNode = oNode.nextSibling ;
141	}
142
143	if ( iCount == 0 )
144	{
145		if ( isBlockElement && FCKConfig.FillEmptyBlocks )
146		{
147			this._AppendEntity( xmlNode, this._NbspEntity ) ;
148			return xmlNode ;
149		}
150
151		var sNodeName = xmlNode.nodeName ;
152
153		// Some inline elements are required to have something inside (span, strong, etc...).
154		if ( FCKListsLib.InlineChildReqElements[ sNodeName ] )
155			return null ;
156
157		// We can't use short representation of empty elements that are not marked
158		// as empty in th XHTML DTD.
159		if ( !FCKListsLib.EmptyElements[ sNodeName ] )
160			xmlNode.appendChild( this.XML.createTextNode('') ) ;
161	}
162
163	return xmlNode ;
164}
165
166FCKXHtml._AppendNode = function( xmlNode, htmlNode )
167{
168	if ( !htmlNode )
169		return false ;
170
171	switch ( htmlNode.nodeType )
172	{
173		// Element Node.
174		case 1 :
175
176			// Here we found an element that is not the real element, but a
177			// fake one (like the Flash placeholder image), so we must get the real one.
178			if ( htmlNode.getAttribute('_fckfakelement') )
179				return FCKXHtml._AppendNode( xmlNode, FCK.GetRealElement( htmlNode ) ) ;
180
181			// Mozilla insert custom nodes in the DOM.
182			if ( FCKBrowserInfo.IsGecko && htmlNode.hasAttribute('_moz_editor_bogus_node') )
183				return false ;
184
185			// This is for elements that are instrumental to FCKeditor and
186			// must be removed from the final HTML.
187			if ( htmlNode.getAttribute('_fcktemp') )
188				return false ;
189
190			// Get the element name.
191			var sNodeName = htmlNode.tagName.toLowerCase()  ;
192
193			if ( FCKBrowserInfo.IsIE )
194			{
195				// IE doens't include the scope name in the nodeName. So, add the namespace.
196				if ( htmlNode.scopeName && htmlNode.scopeName != 'HTML' && htmlNode.scopeName != 'FCK' )
197					sNodeName = htmlNode.scopeName.toLowerCase() + ':' + sNodeName ;
198			}
199			else
200			{
201				if ( sNodeName.StartsWith( 'fck:' ) )
202					sNodeName = sNodeName.Remove( 0,4 ) ;
203			}
204
205			// Check if the node name is valid, otherwise ignore this tag.
206			// If the nodeName starts with a slash, it is a orphan closing tag.
207			// On some strange cases, the nodeName is empty, even if the node exists.
208			if ( !FCKRegexLib.ElementName.test( sNodeName ) )
209				return false ;
210
211			// Remove the <br> if it is a bogus node.
212			if ( sNodeName == 'br' && htmlNode.getAttribute( 'type', 2 ) == '_moz' )
213				return false ;
214
215			// The already processed nodes must be marked to avoid then to be duplicated (bad formatted HTML).
216			// So here, the "mark" is checked... if the element is Ok, then mark it.
217			if ( htmlNode._fckxhtmljob && htmlNode._fckxhtmljob == FCKXHtml.CurrentJobNum )
218				return false ;
219
220			var oNode = this._CreateNode( sNodeName ) ;
221
222			// Add all attributes.
223			FCKXHtml._AppendAttributes( xmlNode, htmlNode, oNode, sNodeName ) ;
224
225			htmlNode._fckxhtmljob = FCKXHtml.CurrentJobNum ;
226
227			// Tag specific processing.
228			var oTagProcessor = FCKXHtml.TagProcessors[ sNodeName ] ;
229
230			if ( oTagProcessor )
231				oNode = oTagProcessor( oNode, htmlNode, xmlNode ) ;
232			else
233				oNode = this._AppendChildNodes( oNode, htmlNode, Boolean( FCKListsLib.NonEmptyBlockElements[ sNodeName ] ) ) ;
234
235			if ( !oNode )
236				return false ;
237
238			xmlNode.appendChild( oNode ) ;
239
240			break ;
241
242		// Text Node.
243		case 3 :
244			return this._AppendTextNode( xmlNode, htmlNode.nodeValue.ReplaceNewLineChars(' ') ) ;
245
246		// Comment
247		case 8 :
248			// IE catches the <!DOTYPE ... > as a comment, but it has no
249			// innerHTML, so we can catch it, and ignore it.
250			if ( FCKBrowserInfo.IsIE && !htmlNode.innerHTML )
251				break ;
252
253			try { xmlNode.appendChild( this.XML.createComment( htmlNode.nodeValue ) ) ; }
254			catch (e) { /* Do nothing... probably this is a wrong format comment. */ }
255			break ;
256
257		// Unknown Node type.
258		default :
259			xmlNode.appendChild( this.XML.createComment( "Element not supported - Type: " + htmlNode.nodeType + " Name: " + htmlNode.nodeName ) ) ;
260			break ;
261	}
262	return true ;
263}
264
265function FCKXHtml_CreateNode_StrongEm( nodeName )
266{
267	switch ( nodeName )
268	{
269		case 'b' :
270			nodeName = 'strong' ;
271			break ;
272		case 'i' :
273			nodeName = 'em' ;
274			break ;
275	}
276	return this.XML.createElement( nodeName ) ;
277}
278
279function FCKXHtml_CreateNode_Normal( nodeName )
280{
281	return this.XML.createElement( nodeName ) ;
282}
283
284// Append an item to the SpecialBlocks array and returns the tag to be used.
285FCKXHtml._AppendSpecialItem = function( item )
286{
287	return '___FCKsi___' + FCKXHtml.SpecialBlocks.AddItem( item ) ;
288}
289
290FCKXHtml._AppendEntity = function( xmlNode, entity )
291{
292	xmlNode.appendChild( this.XML.createTextNode( '#?-:' + entity + ';' ) ) ;
293}
294
295FCKXHtml._AppendTextNode = function( targetNode, textValue )
296{
297	var bHadText = textValue.length > 0 ;
298	if ( bHadText )
299		targetNode.appendChild( this.XML.createTextNode( textValue.replace( FCKXHtmlEntities.EntitiesRegex, FCKXHtml_GetEntity ) ) ) ;
300	return bHadText ;
301}
302
303// Retrieves a entity (internal format) for a given character.
304function FCKXHtml_GetEntity( character )
305{
306	// We cannot simply place the entities in the text, because the XML parser
307	// will translate & to &amp;. So we use a temporary marker which is replaced
308	// in the end of the processing.
309	var sEntity = FCKXHtmlEntities.Entities[ character ] || ( '#' + character.charCodeAt(0) ) ;
310	return '#?-:' + sEntity + ';' ;
311}
312
313// Remove part of an attribute from a node according to a regExp
314FCKXHtml._RemoveAttribute = function( xmlNode, regX, sAttribute )
315{
316	var oAtt = xmlNode.attributes.getNamedItem( sAttribute ) ;
317
318	if ( oAtt && regX.test( oAtt.nodeValue ) )
319	{
320		var sValue = oAtt.nodeValue.replace( regX, '' ) ;
321
322		if ( sValue.length == 0 )
323			xmlNode.attributes.removeNamedItem( sAttribute ) ;
324		else
325			oAtt.nodeValue = sValue ;
326	}
327}
328
329// An object that hold tag specific operations.
330FCKXHtml.TagProcessors =
331{
332	img : function( node, htmlNode )
333	{
334		// The "ALT" attribute is required in XHTML.
335		if ( ! node.attributes.getNamedItem( 'alt' ) )
336			FCKXHtml._AppendAttribute( node, 'alt', '' ) ;
337
338		var sSavedUrl = htmlNode.getAttribute( '_fcksavedurl' ) ;
339		if ( sSavedUrl != null )
340			FCKXHtml._AppendAttribute( node, 'src', sSavedUrl ) ;
341
342		return node ;
343	},
344
345	a : function( node, htmlNode )
346	{
347		// Firefox may create empty tags when deleting the selection in some special cases (SF-BUG 1556878).
348		if ( htmlNode.innerHTML.Trim().length == 0 && !htmlNode.name )
349			return false ;
350
351		var sSavedUrl = htmlNode.getAttribute( '_fcksavedurl' ) ;
352		if ( sSavedUrl != null )
353			FCKXHtml._AppendAttribute( node, 'href', sSavedUrl ) ;
354
355
356		// Anchors with content has been marked with an additional class, now we must remove it.
357		if ( FCKBrowserInfo.IsIE )
358		{
359			FCKXHtml._RemoveAttribute( node, FCKRegexLib.FCK_Class, 'class' ) ;
360
361			// Buggy IE, doesn't copy the name of changed anchors.
362			if ( htmlNode.name )
363				FCKXHtml._AppendAttribute( node, 'name', htmlNode.name ) ;
364		}
365
366		node = FCKXHtml._AppendChildNodes( node, htmlNode, false ) ;
367
368		return node ;
369	},
370
371	script : function( node, htmlNode )
372	{
373		// The "TYPE" attribute is required in XHTML.
374		if ( ! node.attributes.getNamedItem( 'type' ) )
375			FCKXHtml._AppendAttribute( node, 'type', 'text/javascript' ) ;
376
377		node.appendChild( FCKXHtml.XML.createTextNode( FCKXHtml._AppendSpecialItem( htmlNode.text ) ) ) ;
378
379		return node ;
380	},
381
382	style : function( node, htmlNode )
383	{
384		// The "TYPE" attribute is required in XHTML.
385		if ( ! node.attributes.getNamedItem( 'type' ) )
386			FCKXHtml._AppendAttribute( node, 'type', 'text/css' ) ;
387
388		node.appendChild( FCKXHtml.XML.createTextNode( FCKXHtml._AppendSpecialItem( htmlNode.innerHTML ) ) ) ;
389
390		return node ;
391	},
392
393	title : function( node, htmlNode )
394	{
395		node.appendChild( FCKXHtml.XML.createTextNode( FCK.EditorDocument.title ) ) ;
396
397		return node ;
398	},
399
400	table : function( node, htmlNode )
401	{
402		// There is a trick to show table borders when border=0. We add to the
403		// table class the FCK__ShowTableBorders rule. So now we must remove it.
404
405		if ( FCKBrowserInfo.IsIE )
406			FCKXHtml._RemoveAttribute( node, FCKRegexLib.FCK_Class, 'class' ) ;
407
408		node = FCKXHtml._AppendChildNodes( node, htmlNode, false ) ;
409
410		return node ;
411	},
412
413	// Fix nested <ul> and <ol>.
414	ol : function( node, htmlNode, targetNode )
415	{
416		if ( htmlNode.innerHTML.Trim().length == 0 )
417			return false ;
418
419		var ePSibling = targetNode.lastChild ;
420
421		if ( ePSibling && ePSibling.nodeType == 3 )
422			ePSibling = ePSibling.previousSibling ;
423
424		if ( ePSibling && ePSibling.nodeName.toUpperCase() == 'LI' )
425		{
426			htmlNode._fckxhtmljob = null ;
427			FCKXHtml._AppendNode( ePSibling, htmlNode ) ;
428			return false ;
429		}
430
431		node = FCKXHtml._AppendChildNodes( node, htmlNode ) ;
432
433		return node ;
434	},
435
436	span : function( node, htmlNode )
437	{
438		// Firefox may create empty tags when deleting the selection in some special cases (SF-BUG 1084404).
439		if ( htmlNode.innerHTML.length == 0 )
440			return false ;
441
442		node = FCKXHtml._AppendChildNodes( node, htmlNode, false ) ;
443
444		return node ;
445	},
446
447	// IE loses contents of iframes, and Gecko does give it back HtmlEncoded
448	// Note: Opera does lose the content and doesn't provide it in the innerHTML string
449	iframe : function( node, htmlNode )
450	{
451		var sHtml = htmlNode.innerHTML ;
452
453		// Gecko does give back the encoded html
454		if ( FCKBrowserInfo.IsGecko )
455			sHtml = FCKTools.HTMLDecode( sHtml );
456
457		// Remove the saved urls here as the data won't be processed as nodes
458		sHtml = sHtml.replace( /\s_fcksavedurl="[^"]*"/g, '' ) ;
459
460		node.appendChild( FCKXHtml.XML.createTextNode( FCKXHtml._AppendSpecialItem( sHtml ) ) ) ;
461
462		return node ;
463	}
464} ;
465
466FCKXHtml.TagProcessors.ul = FCKXHtml.TagProcessors.ol ;
467