// Modified showdown to only parse a few things

function strip_tags(str, allowed_tags) {
    // Strips HTML and PHP tags from a string  
    // 
    // version: 1003.2411
    // discuss at: http://phpjs.org/functions/strip_tags    // +   original by: Kevin van Zonneveld (http://kevin.vanzonneveld.net)
    // +   improved by: Luke Godfrey
    // +      input by: Pul
    // +   bugfixed by: Kevin van Zonneveld (http://kevin.vanzonneveld.net)
    // +   bugfixed by: Onno Marsman    // +      input by: Alex
    // +   bugfixed by: Kevin van Zonneveld (http://kevin.vanzonneveld.net)
    // +      input by: Marc Palau
    // +   improved by: Kevin van Zonneveld (http://kevin.vanzonneveld.net)
    // +      input by: Brett Zamir (http://brett-zamir.me)    // +   bugfixed by: Kevin van Zonneveld (http://kevin.vanzonneveld.net)
    // +   bugfixed by: Eric Nagel
    // +      input by: Bobby Drake
    // +   bugfixed by: Kevin van Zonneveld (http://kevin.vanzonneveld.net)
    // +   bugfixed by: Tomasz Wesolowski    // *     example 1: strip_tags('<p>Kevin</p> <b>van</b> <i>Zonneveld</i>', '<i><b>');
    // *     returns 1: 'Kevin <b>van</b> <i>Zonneveld</i>'
    // *     example 2: strip_tags('<p>Kevin <img src="someimage.png" onmouseover="someFunction()">van <i>Zonneveld</i></p>', '<p>');
    // *     returns 2: '<p>Kevin van Zonneveld</p>'
    // *     example 3: strip_tags("<a href='http://kevin.vanzonneveld.net'>Kevin van Zonneveld</a>", "<a>");    // *     returns 3: '<a href='http://kevin.vanzonneveld.net'>Kevin van Zonneveld</a>'
    // *     example 4: strip_tags('1 < 5 5 > 1');
    // *     returns 4: '1 < 5 5 > 1'
    var key = '', allowed = false;
    var matches = [];    var allowed_array = [];
    var allowed_tag = '';
    var i = 0;
    var k = '';
    var html = ''; 
    var replacer = function (search, replace, str) {
        return str.split(search).join(replace);
    };
     // Build allowes tags associative array
    if (allowed_tags) {
        allowed_array = allowed_tags.match(/([a-zA-Z0-9]+)/gi);
    }
     str += '';
 
    // Match tags
    matches = str.match(/(<\/?[\S][^>]*>)/gi);
     // Go through all HTML tags
    for (key in matches) {
        if (isNaN(key)) {
            // IE7 Hack
            continue;        }
 
        // Save HTML tag
        html = matches[key].toString();
         // Is tag not in allowed list? Remove from str!
        allowed = false;
 
        // Go through all allowed tags
        for (k in allowed_array) {            // Init
            allowed_tag = allowed_array[k];
            i = -1;
 
            if (i != 0) { i = html.toLowerCase().indexOf('<'+allowed_tag+'>');}
			if (i != 0) { i = html.toLowerCase().indexOf('<'+allowed_tag+' ');}
            if (i != 0) { i = html.toLowerCase().indexOf('</'+allowed_tag);}
 
            // Determine
            if (i == 0) { allowed = true;
                break;
            }
        }
         if (!allowed) {
            str = replacer(html, "", str); // Custom replace. No regexing
        }
    }
     return str;
}

// namespace
var SimpleMarkdown = {};

//
// converter
//
// Wraps all "globals" so that the only thing
// exposed is makeHtml().
//
SimpleMarkdown.converter = function() {

//
// Globals:
//

// Global hashes, used by various utility routines
var g_urls;
var g_titles;
var g_html_blocks;

this.makeHtml = function(text) {
//
// Main function. The order in which other subs are called here is
// essential. Link and image substitutions need to happen before
// _EscapeSpecialCharsWithinTagAttributes(), so that any *'s or _'s in the <a>
// and <img> tags get encoded.
//

	// Clear the global hashes. If we don't clear these, you get conflicts
	// from other articles when generating a page which contains more than
	// one article (e.g. an index page that shows the N most recent
	// articles):
	g_urls = new Array();
	g_titles = new Array();
	g_html_blocks = new Array();

	// attacklab: Replace ~ with ~T
	// This lets us use tilde as an escape char to avoid md5 hashes
	// The choice of character is arbitray; anything that isn't
    // magic in Markdown will work.
	text = text.replace(/~/g,"~T");

	// attacklab: Replace $ with ~D
	// RegExp interprets $ as a special character
	// when it's in a replacement string
	text = text.replace(/\$/g,"~D");

	// Standardize line endings
	text = text.replace(/\r\n/g,"\n"); // DOS to Unix
	text = text.replace(/\r/g,"\n"); // Mac to Unix

	// Make sure text begins and ends with a couple of newlines:
	text = "\n\n" + text + "\n\n";

	// Convert all tabs to spaces.
	text = _Detab(text);

	// Strip any lines consisting only of spaces and tabs.
	// This makes subsequent regexen easier to write, because we can
	// match consecutive blank lines with /\n+/ instead of something
	// contorted like /[ \t]*\n+/ .
	text = text.replace(/^[ \t]+$/mg,"");

	// Turn block-level HTML blocks into hash entries
	text = _HashHTMLBlocks(text);

	// Strip link definitions, store in hashes.
	text = _StripLinkDefinitions(text);

	text = _RunBlockGamut(text);

	text = _UnescapeSpecialChars(text);

	// attacklab: Restore dollar signs
	text = text.replace(/~D/g,"$$");

	// attacklab: Restore tildes
	text = text.replace(/~T/g,"~");

	return text;
}


var _StripLinkDefinitions = function(text) {
//
// Strips link definitions from text, stores the URLs and titles in
// hash references.
//

	// Link defs are in the form: ^[id]: url "optional title"

	/*
		var text = text.replace(/
				^[ ]{0,3}\[(.+)\]:  // id = $1  attacklab: g_tab_width - 1
				  [ \t]*
				  \n?				// maybe *one* newline
				  [ \t]*
				<?(\S+?)>?			// url = $2
				  [ \t]*
				  \n?				// maybe one newline
				  [ \t]*
				(?:
				  (\n*)				// any lines skipped = $3 attacklab: lookbehind removed
				  ["(]
				  (.+?)				// title = $4
				  [")]
				  [ \t]*
				)?					// title is optional
				(?:\n+|$)
			  /gm,
			  function(){...});
	*/
	var text = text.replace(/^[ ]{0,3}\[(.+)\]:[ \t]*\n?[ \t]*<?(\S+?)>?[ \t]*\n?[ \t]*(?:(\n*)["(](.+?)[")][ \t]*)?(?:\n+|\Z)/gm,
		function (wholeMatch,m1,m2,m3,m4) {
			m1 = m1.toLowerCase();
			g_urls[m1] = _EncodeAmpsAndAngles(m2);  // Link IDs are case-insensitive
			if (m3) {
				// Oops, found blank lines, so it's not a title.
				// Put back the parenthetical statement we stole.
				return m3+m4;
			} else if (m4) {
				g_titles[m1] = m4.replace(/"/g,"&quot;");
			}
			
			// Completely remove the definition from the text
			return "";
		}
	);

	return text;
}


var _HashHTMLBlocks = function(text) {
	// attacklab: Double up blank lines to reduce lookaround
	text = text.replace(/\n/g,"\n\n");

	// Hashify HTML blocks:
	// We only want to do this for block-level HTML tags, such as headers,
	// lists, and tables. That's because we still want to wrap <p>s around
	// "paragraphs" that are wrapped in non-block-level tags, such as anchors,
	// phrase emphasis, and spans. The list of tags we're looking for is
	// hard-coded:
	var block_tags_a = "p|pre|table|dl|ol|ul|script|noscript|form|fieldset|iframe|math|ins|del"
	var block_tags_b = "p|pre|table|dl|ol|ul|script|noscript|form|fieldset|iframe|math"

	// First, look for nested blocks, e.g.:
	//   <div>
	//     <div>
	//     tags for inner block must be indented.
	//     </div>
	//   </div>
	//
	// The outermost tags must start at the left margin for this to match, and
	// the inner nested divs must be indented.
	// We need to do this before the next, more liberal match, because the next
	// match will start at the first `<div>` and stop at the first `</div>`.

	// attacklab: This regex can be expensive when it fails.
	/*
		var text = text.replace(/
		(						// save in $1
			^					// start of line  (with /m)
			<($block_tags_a)	// start tag = $2
			\b					// word break
								// attacklab: hack around khtml/pcre bug...
			[^\r]*?\n			// any number of lines, minimally matching
			</\2>				// the matching end tag
			[ \t]*				// trailing spaces/tabs
			(?=\n+)				// followed by a newline
		)						// attacklab: there are sentinel newlines at end of document
		/gm,function(){...}};
	*/
	text = text.replace(/^(<(p|div|h[1-6]|blockquote|pre|table|dl|ol|ul|script|noscript|form|fieldset|iframe|math|ins|del)\b[^\r]*?\n<\/\2>[ \t]*(?=\n+))/gm,hashElement);

	//
	// Now match more liberally, simply from `\n<tag>` to `</tag>\n`
	//

	/*
		var text = text.replace(/
		(						// save in $1
			^					// start of line  (with /m)
			<($block_tags_b)	// start tag = $2
			\b					// word break
								// attacklab: hack around khtml/pcre bug...
			[^\r]*?				// any number of lines, minimally matching
			.*</\2>				// the matching end tag
			[ \t]*				// trailing spaces/tabs
			(?=\n+)				// followed by a newline
		)						// attacklab: there are sentinel newlines at end of document
		/gm,function(){...}};
	*/
	text = text.replace(/^(<(p|div|h[1-6]|blockquote|pre|table|dl|ol|ul|script|noscript|form|fieldset|iframe|math)\b[^\r]*?.*<\/\2>[ \t]*(?=\n+)\n)/gm,hashElement);

	// Special case just for <hr />. It was easier to make a special case than
	// to make the other regex more complicated.  

	/*
		text = text.replace(/
		(						// save in $1
			\n\n				// Starting after a blank line
			[ ]{0,3}
			(<(hr)				// start tag = $2
			\b					// word break
			([^<>])*?			// 
			\/?>)				// the matching end tag
			[ \t]*
			(?=\n{2,})			// followed by a blank line
		)
		/g,hashElement);
	*/
	text = text.replace(/(\n[ ]{0,3}(<(hr)\b([^<>])*?\/?>)[ \t]*(?=\n{2,}))/g,hashElement);

	// Special case for standalone HTML comments:

	/*
		text = text.replace(/
		(						// save in $1
			\n\n				// Starting after a blank line
			[ ]{0,3}			// attacklab: g_tab_width - 1
			<!
			(--[^\r]*?--\s*)+
			>
			[ \t]*
			(?=\n{2,})			// followed by a blank line
		)
		/g,hashElement);
	*/
	text = text.replace(/(\n\n[ ]{0,3}<!(--[^\r]*?--\s*)+>[ \t]*(?=\n{2,}))/g,hashElement);

	// PHP and ASP-style processor instructions (<?...?> and <%...%>)

	/*
		text = text.replace(/
		(?:
			\n\n				// Starting after a blank line
		)
		(						// save in $1
			[ ]{0,3}			// attacklab: g_tab_width - 1
			(?:
				<([?%])			// $2
				[^\r]*?
				\2>
			)
			[ \t]*
			(?=\n{2,})			// followed by a blank line
		)
		/g,hashElement);
	*/
	text = text.replace(/(?:\n\n)([ ]{0,3}(?:<([?%])[^\r]*?\2>)[ \t]*(?=\n{2,}))/g,hashElement);

	// attacklab: Undo double lines (see comment at top of this function)
	text = text.replace(/\n\n/g,"\n");
	return text;
}

var hashElement = function(wholeMatch,m1) {
	var blockText = m1;

	// Undo double lines
	blockText = blockText.replace(/\n\n/g,"\n");
	blockText = blockText.replace(/^\n/,"");
	
	// strip trailing blank lines
	blockText = blockText.replace(/\n+$/g,"");
	
	// Replace the element text with a marker ("~KxK" where x is its key)
	blockText = "\n\n~K" + (g_html_blocks.push(blockText)-1) + "K\n\n";
	
	return blockText;
};

var _RunBlockGamut = function(text) {
	text = _FormParagraphs(text);
	return text;
}


var _RunSpanGamut = function(text) {
	text = _EscapeSpecialCharsWithinTagAttributes(text);
	text = _EncodeBackslashEscapes(text);

	text = _DoAnchors(text);

	// Make links out of things like `<http://example.com/>`
	// Must come after _DoAnchors(), because you can use < and >
	// delimiters in inline links like [this](<url>).
	text = _DoAutoLinks(text);
	text = _EncodeAmpsAndAngles(text);
	text = _DoItalicsAndBold(text);

	// Do hard breaks:
	text = text.replace(/  +\n/g," <br />\n");

	return text;
}

var _EscapeSpecialCharsWithinTagAttributes = function(text) {
//
// Within tags -- meaning between < and > -- encode [\ ` * _] so they
// don't conflict with their use in Markdown for code, italics and strong.
//

	// Build a regex to find HTML tags and comments.  See Friedl's 
	// "Mastering Regular Expressions", 2nd Ed., pp. 200-201.
	var regex = /(<[a-z\/!$]("[^"]*"|'[^']*'|[^'">])*>|<!(--.*?--\s*)+>)/gi;

	text = text.replace(regex, function(wholeMatch) {
		var tag = wholeMatch.replace(/(.)<\/?code>(?=.)/g,"$1`");
		tag = escapeCharacters(tag,"\\`*_");
		return tag;
	});

	return text;
}

var _DoAnchors = function(text) {
//
// Turn Markdown link shortcuts into XHTML <a> tags.
//
	//
	// First, handle reference-style links: [link text] [id]
	//

	/*
		text = text.replace(/
		(							// wrap whole match in $1
			\[
			(
				(?:
					\[[^\]]*\]		// allow brackets nested one level
					|
					[^\[]			// or anything else
				)*
			)
			\]

			[ ]?					// one optional space
			(?:\n[ ]*)?				// one optional newline followed by spaces

			\[
			(.*?)					// id = $3
			\]
		)()()()()					// pad remaining backreferences
		/g,_DoAnchors_callback);
	*/
	text = text.replace(/(\[((?:\[[^\]]*\]|[^\[\]])*)\][ ]?(?:\n[ ]*)?\[(.*?)\])()()()()/g,writeAnchorTag);

	//
	// Next, inline-style links: [link text](url "optional title")
	//

	/*
		text = text.replace(/
			(						// wrap whole match in $1
				\[
				(
					(?:
						\[[^\]]*\]	// allow brackets nested one level
					|
					[^\[\]]			// or anything else
				)
			)
			\]
			\(						// literal paren
			[ \t]*
			()						// no id, so leave $3 empty
			<?(.*?)>?				// href = $4
			[ \t]*
			(						// $5
				(['"])				// quote char = $6
				(.*?)				// Title = $7
				\6					// matching quote
				[ \t]*				// ignore any spaces/tabs between closing quote and )
			)?						// title is optional
			\)
		)
		/g,writeAnchorTag);
	*/
	text = text.replace(/(\[((?:\[[^\]]*\]|[^\[\]])*)\]\([ \t]*()<?(.*?)>?[ \t]*((['"])(.*?)\6[ \t]*)?\))/g,writeAnchorTag);

	//
	// Last, handle reference-style shortcuts: [link text]
	// These must come last in case you've also got [link test][1]
	// or [link test](/foo)
	//

	/*
		text = text.replace(/
		(		 					// wrap whole match in $1
			\[
			([^\[\]]+)				// link text = $2; can't contain '[' or ']'
			\]
		)()()()()()					// pad rest of backreferences
		/g, writeAnchorTag);
	*/
	text = text.replace(/(\[([^\[\]]+)\])()()()()()/g, writeAnchorTag);

	return text;
}

var writeAnchorTag = function(wholeMatch,m1,m2,m3,m4,m5,m6,m7) {
	if (m7 == undefined) m7 = "";
	var whole_match = m1;
	var link_text   = m2;
	var link_id	 = m3.toLowerCase();
	var url		= m4;
	var title	= m7;
	
	if (url == "") {
		if (link_id == "") {
			// lower-case and turn embedded newlines into spaces
			link_id = link_text.toLowerCase().replace(/ ?\n/g," ");
		}
		url = "#"+link_id;
		
		if (g_urls[link_id] != undefined) {
			url = g_urls[link_id];
			if (g_titles[link_id] != undefined) {
				title = g_titles[link_id];
			}
		}
		else {
			if (whole_match.search(/\(\s*\)$/m)>-1) {
				// Special case for explicit empty url
				url = "";
			} else {
				return whole_match;
			}
		}
	}	
	
	url = escapeCharacters(url,"*_");
	var result = "<a href=\"" + url + "\"";
	
	if (title != "") {
		title = title.replace(/"/g,"&quot;");
		title = escapeCharacters(title,"*_");
		result +=  " title=\"" + title + "\"";
	}
	
	result += ">" + link_text + "</a>";
	
	return result;
}
var _DoItalicsAndBold = function(text) {

	// <strong> must go first:
	text = text.replace(/(\*\*|__)(?=\S)([^\r]*?\S[*_]*)\1/g,
		"<strong>$2</strong>");

	text = text.replace(/(\*|_)(?=\S)([^\r]*?\S)\1/g,
		"<em>$2</em>");

	return text;
}

var _FormParagraphs = function(text) {
//
//  Params:
//    $text - string to process with html <p> tags
//

	// Strip leading and trailing lines:
	text = text.replace(/^\n+/g,"");
	text = text.replace(/\n+$/g,"");

	var grafs = text.split(/\n{2,}/g);
	var grafsOut = new Array();

	//
	// Wrap <p> tags.
	//
	var end = grafs.length;
	for (var i=0; i<end; i++) {
		var str = grafs[i];

		// if this is an HTML marker, copy it
		if (str.search(/~K(\d+)K/g) >= 0) {
			grafsOut.push(str);
		}
		else if (str.search(/\S/) >= 0) {
			str = _RunSpanGamut(str);
			str = str.replace(/^([ \t]*)/g,"<p>");
			str += "</p>"
			grafsOut.push(str);
		}

	}

	//
	// Unhashify HTML blocks
	//
	end = grafsOut.length;
	for (var i=0; i<end; i++) {
		// if this is a marker for an html block...
		while (grafsOut[i].search(/~K(\d+)K/) >= 0) {
			var blockText = g_html_blocks[RegExp.$1];
			blockText = blockText.replace(/\$/g,"$$$$"); // Escape any dollar signs
			grafsOut[i] = grafsOut[i].replace(/~K\d+K/,blockText);
		}
	}

	return grafsOut.join("\n\n");
}


var _EncodeAmpsAndAngles = function(text) {
// Smart processing for ampersands and angle brackets that need to be encoded.
	
	// Ampersand-encoding based entirely on Nat Irons's Amputator MT plugin:
	//   http://bumppo.net/projects/amputator/
	text = text.replace(/&(?!#?[xX]?(?:[0-9a-fA-F]+|\w+);)/g,"&amp;");
	
	// Encode naked <'s
	text = text.replace(/<(?![a-z\/?\$!])/gi,"&lt;");
	
	return text;
}


var _EncodeBackslashEscapes = function(text) {
//
//   Parameter:  String.
//   Returns:	The string, with after processing the following backslash
//			   escape sequences.
//

	// attacklab: The polite way to do this is with the new
	// escapeCharacters() function:
	//
	// 	text = escapeCharacters(text,"\\",true);
	// 	text = escapeCharacters(text,"`*_{}[]()>#+-.!",true);
	//
	// ...but we're sidestepping its use of the (slow) RegExp constructor
	// as an optimization for Firefox.  This function gets called a LOT.

	text = text.replace(/\\(\\)/g,escapeCharacters_callback);
	text = text.replace(/\\([`*_{}\[\]()>#+-.!])/g,escapeCharacters_callback);
	return text;
}


var _DoAutoLinks = function(text) {

	text = text.replace(/<((https?|ftp|dict):[^'">\s]+)>/gi,"<a href=\"$1\">$1</a>");

	// Email addresses: <address@domain.foo>

	/*
		text = text.replace(/
			<
			(?:mailto:)?
			(
				[-.\w]+
				\@
				[-a-z0-9]+(\.[-a-z0-9]+)*\.[a-z]+
			)
			>
		/gi, _DoAutoLinks_callback());
	*/
	text = text.replace(/<(?:mailto:)?([-.\w]+\@[-a-z0-9]+(\.[-a-z0-9]+)*\.[a-z]+)>/gi,
		function(wholeMatch,m1) {
			return _EncodeEmailAddress( _UnescapeSpecialChars(m1) );
		}
	);

	return text;
}


var _EncodeEmailAddress = function(addr) {
//
//  Input: an email address, e.g. "foo@example.com"
//
//  Output: the email address as a mailto link, with each character
//	of the address encoded as either a decimal or hex entity, in
//	the hopes of foiling most address harvesting spam bots. E.g.:
//
//	<a href="&#x6D;&#97;&#105;&#108;&#x74;&#111;:&#102;&#111;&#111;&#64;&#101;
//	   x&#x61;&#109;&#x70;&#108;&#x65;&#x2E;&#99;&#111;&#109;">&#102;&#111;&#111;
//	   &#64;&#101;x&#x61;&#109;&#x70;&#108;&#x65;&#x2E;&#99;&#111;&#109;</a>
//
//  Based on a filter by Matthew Wickline, posted to the BBEdit-Talk
//  mailing list: <http://tinyurl.com/yu7ue>
//

	// attacklab: why can't javascript speak hex?
	function char2hex(ch) {
		var hexDigits = '0123456789ABCDEF';
		var dec = ch.charCodeAt(0);
		return(hexDigits.charAt(dec>>4) + hexDigits.charAt(dec&15));
	}

	var encode = [
		function(ch){return "&#"+ch.charCodeAt(0)+";";},
		function(ch){return "&#x"+char2hex(ch)+";";},
		function(ch){return ch;}
	];

	addr = "mailto:" + addr;

	addr = addr.replace(/./g, function(ch) {
		if (ch == "@") {
		   	// this *must* be encoded. I insist.
			ch = encode[Math.floor(Math.random()*2)](ch);
		} else if (ch !=":") {
			// leave ':' alone (to spot mailto: later)
			var r = Math.random();
			// roughly 10% raw, 45% hex, 45% dec
			ch =  (
					r > .9  ?	encode[2](ch)   :
					r > .45 ?	encode[1](ch)   :
								encode[0](ch)
				);
		}
		return ch;
	});

	addr = "<a href=\"" + addr + "\">" + addr + "</a>";
	addr = addr.replace(/">.+:/g,"\">"); // strip the mailto: from the visible part

	return addr;
}


var _UnescapeSpecialChars = function(text) {
//
// Swap back in all the special characters we've hidden.
//
	text = text.replace(/~E(\d+)E/g,
		function(wholeMatch,m1) {
			var charCodeToReplace = parseInt(m1);
			return String.fromCharCode(charCodeToReplace);
		}
	);
	return text;
}

var _Detab = function(text) {
// attacklab: Detab's completely rewritten for speed.
// In perl we could fix it by anchoring the regexp with \G.
// In javascript we're less fortunate.

	// expand first n-1 tabs
	text = text.replace(/\t(?=\t)/g,"    "); // attacklab: g_tab_width

	// replace the nth with two sentinels
	text = text.replace(/\t/g,"~A~B");

	// use the sentinel to anchor our regex so it doesn't explode
	text = text.replace(/~B(.+?)~A/g,
		function(wholeMatch,m1,m2) {
			var leadingText = m1;
			var numSpaces = 4 - leadingText.length % 4;  // attacklab: g_tab_width

			// there *must* be a better way to do this:
			for (var i=0; i<numSpaces; i++) leadingText+=" ";

			return leadingText;
		}
	);

	// clean up sentinels
	text = text.replace(/~A/g,"    ");  // attacklab: g_tab_width
	text = text.replace(/~B/g,"");

	return text;
}

var escapeCharacters = function(text, charsToEscape, afterBackslash) {
	// First we have to escape the escape characters so that
	// we can build a character class out of them
	var regexString = "([" + charsToEscape.replace(/([\[\]\\])/g,"\\$1") + "])";

	if (afterBackslash) {
		regexString = "\\\\" + regexString;
	}

	var regex = new RegExp(regexString,"g");
	text = text.replace(regex,escapeCharacters_callback);

	return text;
}

var escapeCharacters_callback = function(wholeMatch,m1) {
	var charCodeToEscape = m1.charCodeAt(0);
	return "~E"+charCodeToEscape+"E";
}

} // end of converter

