Markdown.Converter.js

var Markdown;

if (typeof exports === "object" && typeof require === "function") // we're in a CommonJS (e.g. Node.js) module
    Markdown = exports;
else
    Markdown = {};
    
// The following text is included for historical reasons, but should
// be taken with a pinch of salt; it's not all true anymore.

//
// 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.
//


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

(function () {

    function identity(x) { return x; }
    function returnFalse(x) { return false; }

    function HookCollection() { }

    HookCollection.prototype = {

        chain: function (hookname, func) {
            var original = this[hookname];
            if (!original)
                throw new Error("unknown hook " + hookname);

            if (original === identity)
                this[hookname] = func;
            else
                this[hookname] = function (text) {
                    var args = Array.prototype.slice.call(arguments, 0);
                    args[0] = original.apply(null, args);
                    return func.apply(null, args);
                };
        },
        set: function (hookname, func) {
            if (!this[hookname])
                throw new Error("unknown hook " + hookname);
            this[hookname] = func;
        },
        addNoop: function (hookname) {
            this[hookname] = identity;
        },
        addFalse: function (hookname) {
            this[hookname] = returnFalse;
        }
    };

    Markdown.HookCollection = HookCollection;

    // 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)
    function SaveHash() { }
    SaveHash.prototype = {
        set: function (key, value) {
            this["s_" + key] = value;
        },
        get: function (key) {
            return this["s_" + key];
        }
    };

    Markdown.Converter = function () {
        var pluginHooks = this.hooks = new HookCollection();
        
        // given a URL that was encountered by itself (without markup), should return the link text that's to be given to this link
        pluginHooks.addNoop("plainLinkText");
        
        // called with the orignal text as given to makeHtml. The result of this plugin hook is the actual markdown source that will be cooked
        pluginHooks.addNoop("preConversion");
        
        // called with the text once all normalizations have been completed (tabs to spaces, line endings, etc.), but before any conversions have
        pluginHooks.addNoop("postNormalization");
        
        // Called with the text before / after creating block elements like code blocks and lists. Note that this is called recursively
        // with inner content, e.g. it's called with the full text, and then only with the content of a blockquote. The inner
        // call will receive outdented text.
        pluginHooks.addNoop("preBlockGamut");
        pluginHooks.addNoop("postBlockGamut");
        
        // called with the text of a single block element before / after the span-level conversions (bold, code spans, etc.) have been made
        pluginHooks.addNoop("preSpanGamut");
        pluginHooks.addNoop("postSpanGamut");
        
        // called with the final cooked HTML code. The result of this plugin hook is the actual output of makeHtml
        pluginHooks.addNoop("postConversion");

        //
        // Private state of the converter instance:
        //

        // 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;

        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.
            //

            // This will only happen if makeHtml on the same converter instance is called from a plugin hook.
            // Don't do that.
            if (g_urls)
                throw new Error("Recursive call to converter.makeHtml");
        
            // Create the private state objects.
            g_urls = new SaveHash();
            g_titles = new SaveHash();
            g_html_blocks = [];
            g_list_level = 0;

            text = pluginHooks.preConversion(text);

            // 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, "");
            
            text = pluginHooks.postNormalization(text);

            // 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, "~");

            text = pluginHooks.postConversion(text);

            g_html_blocks = g_titles = g_urls = null;

            return text;
        };

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

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

            /*
            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(){...});
            */

            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;
        }

        function _HashHTMLBlocks(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.

            /*
            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`
            //

            /*
            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 and http://meta.stackoverflow.com/q/95256
                    >
                    [ \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;
        }

        function hashElement(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 blockGamutHookCallback = function (t) { return _RunBlockGamut(t); }

        function _RunBlockGamut(text, doNotUnhash) {
            //
            // These are all the transformations that form block-level
            // tags like paragraphs, headers, and list items.
            //
            
            text = pluginHooks.preBlockGamut(text, blockGamutHookCallback);
            
            text = _DoHeaders(text);

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

            text = _DoLists(text);
            text = _DoCodeBlocks(text);
            text = _DoBlockQuotes(text);
            
            text = pluginHooks.postBlockGamut(text, blockGamutHookCallback);

            // 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;
        }

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

            text = pluginHooks.preSpanGamut(text);
            
            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 = text.replace(/~P/g, "://"); // put in place to prevent autolinking; reset now
            
            text = _EncodeAmpsAndAngles(text);
            text = _DoItalicsAndBold(text);

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

            return text;
        }

        function _EscapeSpecialCharsWithinTagAttributes(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, wholeMatch.charAt(1) == "!" ? "\\`*_/" : "\\`*_"); // also escape slashes in comments to prevent autolinking there -- http://meta.stackoverflow.com/questions/95987
                return tag;
            });

            return text;
        }

        function _DoAnchors(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, writeAnchorTag);
            */
            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)
                            |
                            [^()\s]
                        )*?
                    )>?                
                    [ \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]*()<?((?:\([^)]*\)|[^()\s])*?)>?[ \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;
        }

        function writeAnchorTag(wholeMatch, m1, m2, m3, m4, m5, m6, m7) {
            if (m7 == undefined) m7 = "";
            var whole_match = m1;
            var link_text = m2.replace(/:\/\//g, "~P"); // to prevent auto-linking withing the link. will be converted back after the auto-linker runs
            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 = encodeProblemUrlChars(url);
            url = escapeCharacters(url, "*_");
            var result = "<a href=\"" + url + "\"";

            if (title != "") {
                title = attributeEncode(title);
                title = escapeCharacters(title, "*_");
                result += " title=\"" + title + "\"";
            }

            result += ">" + link_text + "</a>";

            return result;
        }

        function _DoImages(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;
        }
        
        function attributeEncode(text) {
            // unconditionally replace angle brackets here -- what ends up in an attribute (e.g. alt or title)
            // never makes sense to have verbatim HTML in it (and the sanitizer would totally break it)
            return text.replace(/>/g, "&gt;").replace(/</g, "&lt;").replace(/"/g, "&quot;");
        }

        function writeImageTag(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 = escapeCharacters(attributeEncode(alt_text), "*_[]()");
            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 = attributeEncode(title);
            title = escapeCharacters(title, "*_");
            result += " title=\"" + title + "\"";
            //}

            result += " />";

            return result;
        }

        function _DoHeaders(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;
        }

        function _DoLists(text, isInsideParagraphlessListItem) {
            //
            // 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, isInsideParagraphlessListItem);

                    // 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: "[*+-]" };

        function _ProcessListItems(list_str, list_type, isInsideParagraphlessListItem) {
            //
            //  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 showdown, 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), /* isInsideParagraphlessListItem= */ true);
                        item = item.replace(/\n$/, ""); // chomp(item)
                        if (!isInsideParagraphlessListItem) // only the outer-most item should run this, otherwise it's run multiple times for the inner ones
                            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;
        }

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

            /*
            text = text.replace(/
                (?:\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|^\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;
        }

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

        function _DoCodeSpans(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);
                    c = c.replace(/:\/\//g, "~P"); // to prevent auto-linking. Not necessary in code *blocks*, but in code spans. Will be converted back after the auto-linker runs.
                    return m1 + "<code>" + c + "</code>";
                }
            );

            return text;
        }

        function _EncodeCode(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;
        }

        function _DoItalicsAndBold(text) {

            // <strong> must go first:
            text = text.replace(/([\W_]|^)(\*\*|__)(?=\S)([^\r]*?\S[\*_]*)\2([\W_]|$)/g,
            "$1<strong>$3</strong>$4");

            text = text.replace(/([\W_]|^)(\*|_)(?=\S)([^\r\*_]*?\S)\2([\W_]|$)/g,
            "$1<em>$3</em>$4");

            return text;
        }

        function _DoBlockQuotes(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;
        }

        function _FormParagraphs(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 = [];
            
            var markerRe = /~K(\d+)K/;

            //
            // 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 (markerRe.test(str)) {
                    grafsOut.push(str);
                }
                else if (/\S/.test(str)) {
                    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++) {
                    var foundAny = true;
                    while (foundAny) { // we may need several runs, since the data may be nested
                        foundAny = false;
                        grafsOut[i] = grafsOut[i].replace(/~K(\d+)K/g, function (wholeMatch, id) {
                            foundAny = true;
                            return g_html_blocks[id];
                        });
                    }
                }
            }
            return grafsOut.join("\n\n");
        }

        function _EncodeAmpsAndAngles(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\/?!]|~D)/gi, "&lt;");

            return text;
        }

        function _EncodeBackslashEscapes(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 charInsideUrl = "[-A-Z0-9+&@#/%?=~_|[\\]()!:,.;]",
            charEndingUrl = "[-A-Z0-9+&@#/%=~_|[\\])]",
            autoLinkRegex = new RegExp("(=\"|<)?\\b(https?|ftp)(://" + charInsideUrl + "*" + charEndingUrl + ")(?=$|\\W)", "gi"),
            endCharRegex = new RegExp(charEndingUrl, "i");

        function handleTrailingParens(wholeMatch, lookbehind, protocol, link) {
            if (lookbehind)
                return wholeMatch;
            if (link.charAt(link.length - 1) !== ")")
                return "<" + protocol + link + ">";
            var parens = link.match(/[()]/g);
            var level = 0;
            for (var i = 0; i < parens.length; i++) {
                if (parens[i] === "(") {
                    if (level <= 0)
                        level = 1;
                    else
                        level++;
                }
                else {
                    level--;
                }
            }
            var tail = "";
            if (level < 0) {
                var re = new RegExp("\\){1," + (-level) + "}$");
                link = link.replace(re, function (trailingParens) {
                    tail = trailingParens;
                    return "";
                });
            }
            if (tail) {
                var lastChar = link.charAt(link.length - 1);
                if (!endCharRegex.test(lastChar)) {
                    tail = lastChar + tail;
                    link = link.substr(0, link.length - 1);
                }
            }
            return "<" + protocol + link + ">" + tail;
        }
        
        function _DoAutoLinks(text) {

            // note that at this point, all other URL in the text are already hyperlinked as <a href=""></a>
            // *except* for the <http://www.foo.com> case

            // automatically add < and > around unadorned raw hyperlinks
            // must be preceded by a non-word character (and not by =" or <) and followed by non-word/EOF character
            // simulating the lookbehind in a consuming way is okay here, since a URL can neither and with a " nor
            // with a <, so there is no risk of overlapping matches.
            text = text.replace(autoLinkRegex, handleTrailingParens);

            //  autolink anything like <http://example.com>
            
            var replacer = function (wholematch, m1) { return "<a href=\"" + m1 + "\">" + pluginHooks.plainLinkText(m1) + "</a>"; }
            text = text.replace(/<((https?|ftp):[^'">\s]+)>/gi, replacer);

            // Email addresses: <address@domain.foo>
            /*
            text = text.replace(/
                <
                (?:mailto:)?
                (
                    [-.\w]+
                    \@
                    [-a-z0-9]+(\.[-a-z0-9]+)*\.[a-z]+
                )
                >
            /gi, _DoAutoLinks_callback());
            */

            /* disabling email autolinking, since we don't do that on the server, either
            text = text.replace(/<(?:mailto:)?([-.\w]+\@[-a-z0-9]+(\.[-a-z0-9]+)*\.[a-z]+)>/gi,
                function(wholeMatch,m1) {
                    return _EncodeEmailAddress( _UnescapeSpecialChars(m1) );
                }
            );
            */
            return text;
        }

        function _UnescapeSpecialChars(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;
        }

        function _Outdent(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;
        }

        function _Detab(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 _problemUrlChars = /(?:["'*()[\]:]|~D)/g;

        // hex-encodes some unusual "problem" chars in URLs to avoid URL detection problems 
        function encodeProblemUrlChars(url) {
            if (!url)
                return "";

            var len = url.length;

            return url.replace(_problemUrlChars, function (match, offset) {
                if (match == "~D") // escape for dollar
                    return "%24";
                if (match == ":") {
                    if (offset == len - 1 || /[0-9\/]/.test(url.charAt(offset + 1)))
                        return ":"
                }
                return "%" + match.charCodeAt(0).toString(16);
            });
        }


        function escapeCharacters(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;
        }


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

    }; // end of the Markdown.Converter constructor

})();