[osqa.git] / forum / skins / default / media / js / wmd / showdown.js

//
// showdown.js -- A javascript port of Markdown.
//
// Copyright (c) 2007 John Fraser.
//
// Original Markdown Copyright (c) 2004-2005 John Gruber
//   <http://daringfireball.net/projects/markdown/>
//
// The full source distribution is at:
//
//                              A A L
//                              T C A
//                              T K B
//
//   <http://www.attacklab.net/>
//

//
// Wherever possible, Showdown is a straight, line-by-line port
// of the Perl version of Markdown.
//
// This is not a normal parser design; it's basically just a
// series of string substitutions.  It's hard to read and
// maintain this way,  but keeping Showdown close to the original
// design makes it easier to port new features.
//
// More importantly, Showdown behaves like markdown.pl in most
// edge cases.  So web applications can do client-side preview
// in Javascript, and then build identical HTML on the server.
//
// This port needs the new RegExp functionality of ECMA 262,
// 3rd Edition (i.e. Javascript 1.5).  Most modern web browsers
// should do fine.  Even with the new regular expression features,
// We do a lot of work to emulate Perl's regex functionality.
// The tricky changes in this file mostly have the "attacklab:"
// label.  Major or self-explanatory changes don't.
//
// Smart diff tools like Araxis Merge will be able to match up
// this file with markdown.pl in a useful way.  A little tweaking
// helps: in a copy of markdown.pl, replace "#" with "//" and
// replace "$text" with "text".  Be sure to ignore whitespace
// and line endings.
//


//
// Showdown usage:
//
//   var text = "Markdown *rocks*.";
//
//   var converter = new Attacklab.showdown.converter();
//   var html = converter.makeHtml(text);
//
//   alert(html);
//
// Note: move the sample code to the bottom of this
// file before uncommenting it.
//


//
// Attacklab namespace
//
var Attacklab = Attacklab || {}

//
// Showdown namespace
//
Attacklab.showdown = Attacklab.showdown || {}

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


// g_urls and g_titles allow arbitrary user-entered strings as keys. This
// caused an exception (and hence stopped the rendering) when the user entered
// e.g. [push] or [__proto__]. Adding a prefix to the actual key prevents this
// (since no builtin property starts with "s_"). See
// http://meta.stackoverflow.com/questions/64655/strange-wmd-bug
// (granted, switching from Array() to Object() alone would have left only __proto__
// to be a problem)
var SaveHash = function () {
    this.set = function (key, value) {
        this["s_" + key] = value;
    }
    this.get = function (key) {
        return this["s_" + key];
    }
}

//
// Globals:
//

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

// Used to track when we're inside an ordered or unordered list
// (see _ProcessListItems() for details):
var g_list_level = 0;


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 SaveHash();
    g_titles = new SaveHash();
        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 = $2
                (?=\s|$)            // lookahead for whitespace instead of the lookbehind removed below
                                  [ \t]*
                                  \n?                           // maybe one newline
                                  [ \t]*
                                (                   // (potential) title = $3
                                  (\n*)                         // any lines skipped = $4 attacklab: lookbehind removed
                  [ \t]+
                                  ["(]
                                  (.+?)                         // title = $5
                                  [")]
                                  [ \t]*
                                )?                                      // title is optional
                                (?:\n+|$)
                          /gm,
                          function(){...});
        */
        var text = text.replace(/^[ ]{0,3}\[(.+)\]:[ \t]*\n?[ \t]*<?(\S+?)>?(?=\s|$)[ \t]*\n?[ \t]*((\n*)["(](.+?)[")][ \t]*)?(?:\n+)/gm,
                function (wholeMatch,m1,m2,m3,m4,m5) {
                        m1 = m1.toLowerCase();
                        g_urls.set(m1, _EncodeAmpsAndAngles(m2));  // Link IDs are case-insensitive
                        if (m4) {
                                // Oops, found blank lines, so it's not a title.
                                // Put back the parenthetical statement we stole.
                                return m3;
                        } else if (m5) {
                                g_titles.set(m1, m5.replace(/"/g,"&quot;"));
                        }
                        
                        // Completely remove the definition from the text
                        return "";
                }
        );

        return text;
}

var _HashHTMLBlocks = function(text) {

        // 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|div|h[1-6]|blockquote|pre|table|dl|ol|ul|script|noscript|form|fieldset|iframe|math|ins|del"
        var block_tags_b = "p|div|h[1-6]|blockquote|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(/
                \n                                  // Starting after a blank line
                [ ]{0,3}
                (                                               // save in $1
                        (<(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(/
                \n\n                            // Starting after a blank line
                [ ]{0,3}                        // attacklab: g_tab_width - 1
                (                                               // save in $1
                        <!
                        (--(?:|(?:[^>-]|-[^>])(?:[^-]|-[^-])*)--)    // see http://www.w3.org/TR/html-markup/syntax.html#comments
                        >
                        [ \t]*
                        (?=\n{2,})                      // followed by a blank line
                )
                /g,hashElement);
        */
        text = text.replace(/\n\n[ ]{0,3}(<!(--(?:|(?:[^>-]|-[^>])(?:[^-]|-[^-])*)--)>[ \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);

        return text;
}

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

        // Undo double lines
        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, doNotUnhash) {
//
// These are all the transformations that form block-level
// tags like paragraphs, headers, and list items.
//
        text = _DoHeaders(text);

        // Do Horizontal Rules:
        var key = hashBlock("<hr />");
        text = text.replace(/^[ ]{0,2}([ ]?\*[ ]?){3,}[ \t]*$/gm,key);
        text = text.replace(/^[ ]{0,2}([ ]?-[ ]?){3,}[ \t]*$/gm,key);
        text = text.replace(/^[ ]{0,2}([ ]?_[ ]?){3,}[ \t]*$/gm,key);

        text = _DoLists(text);
        text = _DoCodeBlocks(text);
        text = _DoBlockQuotes(text);

        // We already ran _HashHTMLBlocks() before, in Markdown(), but that
        // was to escape raw HTML in the original Markdown source. This time,
        // we're escaping the markup we've just created, so that we don't wrap
        // <p> tags around block-level tags.
        text = _HashHTMLBlocks(text);
    text = _FormParagraphs(text, doNotUnhash);

        return text;
}


var _RunSpanGamut = function(text) {
//
// These are all the transformations that occur *within* block-level
// tags like paragraphs, headers, and list items.
//

        text = _DoCodeSpans(text);
        text = _EscapeSpecialCharsWithinTagAttributes(text);
        text = _EncodeBackslashEscapes(text);

        // Process anchor and image tags. Images must come first,
        // because ![foo][f] looks like an anchor.
        text = _DoImages(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.
    
    // SE: changed the comment part of the regex

    var regex = /(<[a-z\/!$]("[^"]*"|'[^']*'|[^'">])*>|<!(--(?:|(?:[^>-]|-[^>])(?:[^-]|-[^-])*)--)>)/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
                (?:
                    \([^)]*\)       // allow one level of (correctly nested) parens (think MSDN)
                    |
                    [^()]
                )*?
            )>?                         
                        [ \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.get(link_id) != undefined) {
                        url = g_urls.get(link_id);
                        if (g_titles.get(link_id) != undefined) {
                                title = g_titles.get(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 _DoImages = function(text) {
//
// Turn Markdown image shortcuts into <img> tags.
//

        //
        // First, handle reference-style labeled images: ![alt text][id]
        //

        /*
                text = text.replace(/
                (                                               // wrap whole match in $1
                        !\[
                        (.*?)                           // alt text = $2
                        \]

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

                        \[
                        (.*?)                           // id = $3
                        \]
                )()()()()                               // pad rest of backreferences
                /g,writeImageTag);
        */
        text = text.replace(/(!\[(.*?)\][ ]?(?:\n[ ]*)?\[(.*?)\])()()()()/g,writeImageTag);

        //
        // Next, handle inline images:  ![alt text](url "optional title")
        // Don't forget: encode * and _

        /*
                text = text.replace(/
                (                                               // wrap whole match in $1
                        !\[
                        (.*?)                           // alt text = $2
                        \]
                        \s?                                     // One optional whitespace character
                        \(                                      // literal paren
                        [ \t]*
                        ()                                      // no id, so leave $3 empty
                        <?(\S+?)>?                      // src url = $4
                        [ \t]*
                        (                                       // $5
                                (['"])                  // quote char = $6
                                (.*?)                   // title = $7
                                \6                              // matching quote
                                [ \t]*
                        )?                                      // title is optional
                \)
                )
                /g,writeImageTag);
        */
        text = text.replace(/(!\[(.*?)\]\s?\([ \t]*()<?(\S+?)>?[ \t]*((['"])(.*?)\6[ \t]*)?\))/g,writeImageTag);

        return text;
}

var writeImageTag = function(wholeMatch,m1,m2,m3,m4,m5,m6,m7) {
        var whole_match = m1;
        var alt_text   = m2;
        var link_id      = m3.toLowerCase();
        var url         = m4;
        var title       = m7;

        if (!title) title = "";
        
        if (url == "") {
                if (link_id == "") {
                        // lower-case and turn embedded newlines into spaces
                        link_id = alt_text.toLowerCase().replace(/ ?\n/g," ");
                }
                url = "#"+link_id;
                
                if (g_urls.get(link_id) != undefined) {
                        url = g_urls.get(link_id);
                        if (g_titles.get(link_id) != undefined) {
                                title = g_titles.get(link_id);
                        }
                }
                else {
                        return whole_match;
                }
        }       
        
        alt_text = alt_text.replace(/"/g,"&quot;");
        url = escapeCharacters(url,"*_");
        var result = "<img src=\"" + url + "\" alt=\"" + alt_text + "\"";

        // attacklab: Markdown.pl adds empty title attributes to images.
        // Replicate this bug.

        //if (title != "") {
                title = title.replace(/"/g,"&quot;");
                title = escapeCharacters(title,"*_");
                result +=  " title=\"" + title + "\"";
        //}
        
        result += " />";
        
        return result;
}


var _DoHeaders = function(text) {

        // Setext-style headers:
        //      Header 1
        //      ========
        //  
        //      Header 2
        //      --------
        //
        text = text.replace(/^(.+)[ \t]*\n=+[ \t]*\n+/gm,
                function(wholeMatch,m1){return "<h1>" + _RunSpanGamut(m1) + "</h1>\n\n";});

        text = text.replace(/^(.+)[ \t]*\n-+[ \t]*\n+/gm,
                function(matchFound,m1){return "<h2>" + _RunSpanGamut(m1) + "</h2>\n\n";});

        // atx-style headers:
        //  # Header 1
        //  ## Header 2
        //  ## Header 2 with closing hashes ##
        //  ...
        //  ###### Header 6
        //

        /*
                text = text.replace(/
                        ^(\#{1,6})                              // $1 = string of #'s
                        [ \t]*
                        (.+?)                                   // $2 = Header text
                        [ \t]*
                        \#*                                             // optional closing #'s (not counted)
                        \n+
                /gm, function() {...});
        */

        text = text.replace(/^(\#{1,6})[ \t]*(.+?)[ \t]*\#*\n+/gm,
                function(wholeMatch,m1,m2) {
                        var h_level = m1.length;
                        return "<h" + h_level + ">" + _RunSpanGamut(m2) + "</h" + h_level + ">\n\n";
                });

        return text;
}

// This declaration keeps Dojo compressor from outputting garbage:
var _ProcessListItems;

var _DoLists = function(text) {
//
// Form HTML ordered (numbered) and unordered (bulleted) lists.
//

        // attacklab: add sentinel to hack around khtml/safari bug:
        // http://bugs.webkit.org/show_bug.cgi?id=11231
        text += "~0";

        // Re-usable pattern to match any entirel ul or ol list:

        /*
                var whole_list = /
                (                                                                       // $1 = whole list
                        (                                                               // $2
                                [ ]{0,3}                                        // attacklab: g_tab_width - 1
                                ([*+-]|\d+[.])                          // $3 = first list item marker
                                [ \t]+
                        )
                        [^\r]+?
                        (                                                               // $4
                                ~0                                                      // sentinel for workaround; should be $
                        |
                                \n{2,}
                                (?=\S)
                                (?!                                                     // Negative lookahead for another list item marker
                                        [ \t]*
                                        (?:[*+-]|\d+[.])[ \t]+
                                )
                        )
                )/g
        */
        var whole_list = /^(([ ]{0,3}([*+-]|\d+[.])[ \t]+)[^\r]+?(~0|\n{2,}(?=\S)(?![ \t]*(?:[*+-]|\d+[.])[ \t]+)))/gm;

        if (g_list_level) {
                text = text.replace(whole_list,function(wholeMatch,m1,m2) {
                        var list = m1;
                        var list_type = (m2.search(/[*+-]/g)>-1) ? "ul" : "ol";

                        var result = _ProcessListItems(list, list_type);
        
                        // Trim any trailing whitespace, to put the closing `</$list_type>`
                        // up on the preceding line, to get it past the current stupid
                        // HTML block parser. This is a hack to work around the terrible
                        // hack that is the HTML block parser.
                        result = result.replace(/\s+$/,"");
                        result = "<"+list_type+">" + result + "</"+list_type+">\n";
                        return result;
                });
        } else {
                whole_list = /(\n\n|^\n?)(([ ]{0,3}([*+-]|\d+[.])[ \t]+)[^\r]+?(~0|\n{2,}(?=\S)(?![ \t]*(?:[*+-]|\d+[.])[ \t]+)))/g;
                text = text.replace(whole_list,function(wholeMatch,m1,m2,m3) {
                        var runup = m1;
                        var list = m2;

                        var list_type = (m3.search(/[*+-]/g)>-1) ? "ul" : "ol";
                        var result = _ProcessListItems(list, list_type);
                        result = runup + "<"+list_type+">\n" + result + "</"+list_type+">\n";   
                        return result;
                });
        }

        // attacklab: strip sentinel
        text = text.replace(/~0/,"");

        return text;
}

var _listItemMarkers = { ol: "\\d+[.]", ul: "[*+-]" };

_ProcessListItems = function(list_str, list_type) {
//
//  Process the contents of a single ordered or unordered list, splitting it
//  into individual list items.
//
//  list_type is either "ul" or "ol".

        // The $g_list_level global keeps track of when we're inside a list.
        // Each time we enter a list, we increment it; when we leave a list,
        // we decrement. If it's zero, we're not in a list anymore.
        //
        // We do this because when we're not inside a list, we want to treat
        // something like this:
        //
        //    I recommend upgrading to version
        //    8. Oops, now this line is treated
        //    as a sub-list.
        //
        // As a single paragraph, despite the fact that the second line starts
        // with a digit-period-space sequence.
        //
        // Whereas when we're inside a list (or sub-list), that line will be
        // treated as the start of a sub-list. What a kludge, huh? This is
        // an aspect of Markdown's syntax that's hard to parse perfectly
        // without resorting to mind-reading. Perhaps the solution is to
        // change the syntax rules such that sub-lists must start with a
        // starting cardinal number; e.g. "1." or "a.".

        g_list_level++;

        // trim trailing blank lines:
        list_str = list_str.replace(/\n{2,}$/,"\n");

        // attacklab: add sentinel to emulate \z
        list_str += "~0";

        // In the original attacklab WMD, list_type was not given to this function, and anything
        // that matched /[*+-]|\d+[.]/ would just create the next <li>, causing this mismatch:
        //
    //  Markdown          rendered by WMD        rendered by MarkdownSharp
        //  ------------------------------------------------------------------
        //  1. first          1. first               1. first
        //  2. second         2. second              2. second
        //  - third           3. third                   * third
        //
        // We changed this to behave identical to MarkdownSharp. This is the constructed RegEx,
    // with {MARKER} being one of \d+[.] or [*+-], depending on list_type:
        /*
                list_str = list_str.replace(/
                        (^[ \t]*)                                               // leading whitespace = $1
                        ({MARKER}) [ \t]+                       // list marker = $2
                        ([^\r]+?                                                // list item text   = $3
                        (\n+))
                        (?= (~0 | \2 ({MARKER}) [ \t]+))
                /gm, function(){...});
        */
    
    var marker = _listItemMarkers[list_type];
    var re = new RegExp("(^[ \\t]*)(" + marker + ")[ \\t]+([^\\r]+?(\\n+))(?=(~0|\\1(" + marker + ")[ \\t]+))", "gm");
    var last_item_had_a_double_newline = false;
        list_str = list_str.replace(re,
                function(wholeMatch,m1,m2,m3){
                        var item = m3;
                        var leading_space = m1;
            var ends_with_double_newline = /\n\n$/.test(item);
                        var contains_double_newline = ends_with_double_newline || item.search(/\n{2,}/)>-1;

                        if (contains_double_newline || last_item_had_a_double_newline) {
                                item =  _RunBlockGamut(_Outdent(item), /* doNotUnhash = */ true);
                        }
                        else {
                                // Recursion for sub-lists:
                                item = _DoLists(_Outdent(item));
                                item = item.replace(/\n$/,""); // chomp(item)
                                item = _RunSpanGamut(item);
            }
            last_item_had_a_double_newline = ends_with_double_newline;
                        return  "<li>" + item + "</li>\n";
                }
        );

        // attacklab: strip sentinel
        list_str = list_str.replace(/~0/g,"");

        g_list_level--;
        return list_str;
}


var _DoCodeBlocks = function(text) {
//
//  Process Markdown `<pre><code>` blocks.
//  

        /*
                text = text.replace(text,
                        /(?:\n\n|^)
                        (                                                               // $1 = the code block -- one or more lines, starting with a space/tab
                                (?:
                                        (?:[ ]{4}|\t)                   // Lines must start with a tab or a tab-width of spaces - attacklab: g_tab_width
                                        .*\n+
                                )+
                        )
                        (\n*[ ]{0,3}[^ \t\n]|(?=~0))    // attacklab: g_tab_width
                /g,function(){...});
        */

        // attacklab: sentinel workarounds for lack of \A and \Z, safari\khtml bug
        text += "~0";
        
        text = text.replace(/(?:\n\n|^)((?:(?:[ ]{4}|\t).*\n+)+)(\n*[ ]{0,3}[^ \t\n]|(?=~0))/g,
                function(wholeMatch,m1,m2) {
                        var codeblock = m1;
                        var nextChar = m2;
                
                        codeblock = _EncodeCode( _Outdent(codeblock));
                        codeblock = _Detab(codeblock);
                        codeblock = codeblock.replace(/^\n+/g,""); // trim leading newlines
                        codeblock = codeblock.replace(/\n+$/g,""); // trim trailing whitespace

                        codeblock = "<pre><code>" + codeblock + "\n</code></pre>";

                        return "\n\n" + codeblock + "\n\n" + nextChar;
                }
        );

        // attacklab: strip sentinel
        text = text.replace(/~0/,"");

        return text;
}

var hashBlock = function(text) {
        text = text.replace(/(^\n+|\n+$)/g,"");
        return "\n\n~K" + (g_html_blocks.push(text)-1) + "K\n\n";
}


var _DoCodeSpans = function(text) {
//
//   *  Backtick quotes are used for <code></code> spans.
// 
//   *  You can use multiple backticks as the delimiters if you want to
//       include literal backticks in the code span. So, this input:
//       
//               Just type ``foo `bar` baz`` at the prompt.
//       
//         Will translate to:
//       
//               <p>Just type <code>foo `bar` baz</code> at the prompt.</p>
//       
//      There's no arbitrary limit to the number of backticks you
//      can use as delimters. If you need three consecutive backticks
//      in your code, use four for delimiters, etc.
//
//  *  You can use spaces to get literal backticks at the edges:
//       
//               ... type `` `bar` `` ...
//       
//         Turns to:
//       
//               ... type <code>`bar`</code> ...
//

        /*
                text = text.replace(/
                        (^|[^\\])                                       // Character before opening ` can't be a backslash
                        (`+)                                            // $2 = Opening run of `
                        (                                                       // $3 = The code block
                                [^\r]*?
                                [^`]                                    // attacklab: work around lack of lookbehind
                        )
                        \2                                                      // Matching closer
                        (?!`)
                /gm, function(){...});
        */

        text = text.replace(/(^|[^\\])(`+)([^\r]*?[^`])\2(?!`)/gm,
                function(wholeMatch,m1,m2,m3,m4) {
                        var c = m3;
                        c = c.replace(/^([ \t]*)/g,""); // leading whitespace
                        c = c.replace(/[ \t]*$/g,"");   // trailing whitespace
                        c = _EncodeCode(c);
                        return m1+"<code>"+c+"</code>";
                });

        return text;
}


var _EncodeCode = function(text) {
//
// Encode/escape certain characters inside Markdown code runs.
// The point is that in code, these characters are literals,
// and lose their special Markdown meanings.
//
        // Encode all ampersands; HTML entities are not
        // entities within a Markdown code span.
        text = text.replace(/&/g,"&amp;");

        // Do the angle bracket song and dance:
        text = text.replace(/</g,"&lt;");
        text = text.replace(/>/g,"&gt;");

        // Now, escape characters that are magic in Markdown:
        text = escapeCharacters(text,"\*_{}[]\\",false);

// jj the line above breaks this:
//---

//* Item

//   1. Subitem

//            special char: *
//---

        return text;
}


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 _DoBlockQuotes = function(text) {

        /*
                text = text.replace(/
                (                                                               // Wrap whole match in $1
                        (
                                ^[ \t]*>[ \t]?                  // '>' at the start of a line
                                .+\n                                    // rest of the first line
                                (.+\n)*                                 // subsequent consecutive lines
                                \n*                                             // blanks
                        )+
                )
                /gm, function(){...});
        */

        text = text.replace(/((^[ \t]*>[ \t]?.+\n(.+\n)*\n*)+)/gm,
                function(wholeMatch,m1) {
                        var bq = m1;

                        // attacklab: hack around Konqueror 3.5.4 bug:
                        // "----------bug".replace(/^-/g,"") == "bug"

                        bq = bq.replace(/^[ \t]*>[ \t]?/gm,"~0");       // trim one level of quoting

                        // attacklab: clean up hack
                        bq = bq.replace(/~0/g,"");

                        bq = bq.replace(/^[ \t]+$/gm,"");               // trim whitespace-only lines
                        bq = _RunBlockGamut(bq);                                // recurse
                        
                        bq = bq.replace(/(^|\n)/g,"$1  ");
                        // These leading spaces screw with <pre> content, so we need to fix that:
                        bq = bq.replace(
                                        /(\s*<pre>[^\r]+?<\/pre>)/gm,
                                function(wholeMatch,m1) {
                                        var pre = m1;
                                        // attacklab: hack around Konqueror 3.5.4 bug:
                                        pre = pre.replace(/^  /mg,"~0");
                                        pre = pre.replace(/~0/g,"");
                                        return pre;
                                });
                        
                        return hashBlock("<blockquote>\n" + bq + "\n</blockquote>");
                });
        return text;
}


var _FormParagraphs = function(text, doNotUnhash) {
//
//  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
        //
    if (!doNotUnhash) {
        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 _Outdent = function(text) {
//
// Remove one level of line-leading tabs or spaces
//

        // attacklab: hack around Konqueror 3.5.4 bug:
        // "----------bug".replace(/^-/g,"") == "bug"

        text = text.replace(/^(\t|[ ]{1,4})/gm,"~0"); // attacklab: g_tab_width

        // attacklab: clean up hack
        text = text.replace(/~0/g,"")

        return text;
}

var _Detab = function (text) {
        if (!/\t/.test(text))
                return text;

        var spaces = ["    ", "   ", "  ", " "],
                skew = 0,
                v;

        return text.replace(/[\n\t]/g, function (match, offset) {
                if (match === "\n") {
                        skew = offset + 1;
                        return match;
                }
                v = (offset - skew) % 4;
                skew = offset + 1;
                return spaces[v];
        });
}

//
//  attacklab: Utility functions
//


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 Attacklab.showdown.converter


// Version 0.9 used the Showdown namespace instead of Attacklab.showdown
// The old namespace is deprecated, but we'll support it for now:
var Showdown = Attacklab.showdown;

// If anyone's interested, tell the world that this file's been loaded
if (Attacklab.fileLoaded) {
        Attacklab.fileLoaded("showdown.js");
}