Index: openacs-4/packages/acs-templating/www/resources/xinha-nightly/plugins/SuperClean/filters/word.js
===================================================================
RCS file: /usr/local/cvsroot/openacs-4/packages/acs-templating/www/resources/xinha-nightly/plugins/SuperClean/filters/word.js,v
diff -u -r1.2 -r1.3
--- openacs-4/packages/acs-templating/www/resources/xinha-nightly/plugins/SuperClean/filters/word.js	13 May 2008 18:53:57 -0000	1.2
+++ openacs-4/packages/acs-templating/www/resources/xinha-nightly/plugins/SuperClean/filters/word.js	16 Jan 2009 09:05:49 -0000	1.3
@@ -1,30 +1,53 @@
-/* This compressed file is part of Xinha. For uncompressed sources, forum, and bug reports, go to xinha.org */
-/* This file is part of version 0.95 released Mon, 12 May 2008 17:33:15 +0200 */
-/* The URL of the most recent version of this file is http://svn.xinha.webfactional.com/trunk/plugins/SuperClean/filters/word.js */
-function(_1){
-_1=_1.replace(/<!--[\w\s\d@{}:.;,'"%!#_=&|?~()[*+\/\-\]]*-->/gi,"");
-_1=_1.replace(/<!--[^\0]*-->/gi,"");
-_1=_1.replace(/<\/?\s*HTML[^>]*>/gi,"");
-_1=_1.replace(/<\/?\s*BODY[^>]*>/gi,"");
-_1=_1.replace(/<\/?\s*META[^>]*>/gi,"");
-_1=_1.replace(/<\/?\s*SPAN[^>]*>/gi,"");
-_1=_1.replace(/<\/?\s*FONT[^>]*>/gi,"");
-_1=_1.replace(/<\/?\s*IFRAME[^>]*>/gi,"");
-_1=_1.replace(/<\/?\s*STYLE[^>]*>(.|[\n\r\t])*<\/\s*STYLE\s*>/gi,"");
-_1=_1.replace(/<\s*TITLE[^>]*>(.|[\n\r\t])*<\/\s*TITLE\s*>/gi,"");
-_1=_1.replace(/<\s*SCRIPT[^>]*>[^\0]*<\/\s*SCRIPT\s*>/gi,"");
-_1=_1.replace(/<\s*HEAD[^>]*>(.|[\n\r\t])*<\/\s*HEAD\s*>/gi,"");
-_1=_1.replace(/<\s*(\w[^>]*) class=([^ |>]*)([^>]*)/gi,"<$1$3");
-_1=_1.replace(/<\s*(\w[^>]*) style="([^"]*)"([^>]*)/gi,"<$1$3");
-_1=_1.replace(/<\s*(\w[^>]*) lang=([^ |>]*)([^>]*)/gi,"<$1$3");
-_1=_1.replace(/<\\?\?xml[^>]*>/gi,"");
-_1=_1.replace(/<\/?\w+:[^>]*>/gi,"");
-_1=_1.replace(/&nbsp;/," ");
-_1=_1.replace(/<\s*p[^>]*><\s*br\s*\/?>\s*<\/\s*p[^>]*>/gi,"<br>");
-_1=_1.replace(/<\s*p[^>]*>/gi,"");
-_1=_1.replace(/<\/\s*p[^>]*>/gi,"<br>");
-_1=_1.replace(/(\s*<br>\s*)*$/,"");
-_1=_1.trim();
-return _1;
-}
+function(html) {
+    // Remove HTML comments
+	html = html.replace(/<!--[\w\s\d@{}:.;,'"%!#_=&|?~()[*+\/\-\]]*-->/gi, "" );
+	html = html.replace(/<!--[^\0]*-->/gi, '');
+    // Remove all HTML tags
+	html = html.replace(/<\/?\s*HTML[^>]*>/gi, "" );
+    // Remove all BODY tags
+    html = html.replace(/<\/?\s*BODY[^>]*>/gi, "" );
+    // Remove all META tags
+	html = html.replace(/<\/?\s*META[^>]*>/gi, "" );
+    // Remove all SPAN tags
+	html = html.replace(/<\/?\s*SPAN[^>]*>/gi, "" );
+	// Remove all FONT tags
+    html = html.replace(/<\/?\s*FONT[^>]*>/gi, "");
+    // Remove all IFRAME tags.
+    html = html.replace(/<\/?\s*IFRAME[^>]*>/gi, "");
+    // Remove all STYLE tags & content
+	html = html.replace(/<\/?\s*STYLE[^>]*>(.|[\n\r\t])*<\/\s*STYLE\s*>/gi, "" );
+    // Remove all TITLE tags & content
+	html = html.replace(/<\s*TITLE[^>]*>(.|[\n\r\t])*<\/\s*TITLE\s*>/gi, "" );
+	// Remove javascript
+    html = html.replace(/<\s*SCRIPT[^>]*>[^\0]*<\/\s*SCRIPT\s*>/gi, "");
+    // Remove all HEAD tags & content
+	html = html.replace(/<\s*HEAD[^>]*>(.|[\n\r\t])*<\/\s*HEAD\s*>/gi, "" );
+	// Remove Class attributes
+	html = html.replace(/<\s*(\w[^>]*) class=([^ |>]*)([^>]*)/gi, "<$1$3") ;
+	// Remove Style attributes
+	html = html.replace(/<\s*(\w[^>]*) style="([^"]*)"([^>]*)/gi, "<$1$3") ;
+	// Remove Lang attributes
+	html = html.replace(/<\s*(\w[^>]*) lang=([^ |>]*)([^>]*)/gi, "<$1$3") ;
+	// Remove XML elements and declarations
+	html = html.replace(/<\\?\?xml[^>]*>/gi, "") ;
+	// Remove Tags with XML namespace declarations: <o:p></o:p>
+	html = html.replace(/<\/?\w+:[^>]*>/gi, "") ;
+	// Replace the &nbsp;
+	html = html.replace(/&nbsp;/, " " );
 
+	// Transform <p><br /></p> to <br>
+	//html = html.replace(/<\s*p[^>]*>\s*<\s*br\s*\/>\s*<\/\s*p[^>]*>/gi, "<br>");
+	html = html.replace(/<\s*p[^>]*><\s*br\s*\/?>\s*<\/\s*p[^>]*>/gi, "<br>");
+	
+	// Remove <P> 
+	html = html.replace(/<\s*p[^>]*>/gi, "");
+	
+	// Replace </p> with <br>
+	html = html.replace(/<\/\s*p[^>]*>/gi, "<br>");
+	
+	// Remove any <br> at the end
+	html = html.replace(/(\s*<br>\s*)*$/, "");
+	
+	html = html.trim();
+	return html;
+} 
\ No newline at end of file