discourse/app/assets/javascripts/external/Markdown.Converter.js

1407 lines
53 KiB
JavaScript
Raw Normal View History

/* LICENSE: http://code.google.com/p/pagedown/source/browse/LICENSE.txt */
2013-02-05 14:16:51 -05:00
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);
};
2013-02-05 14:16:51 -05:00
},
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");
2013-02-05 14:16:51 -05:00
//
// 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);
2013-02-05 14:16:51 -05:00
// 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); }
2013-02-05 14:16:51 -05:00
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);
2013-02-05 14:16:51 -05:00
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]*$/gm, replacement);
text = _DoLists(text);
text = _DoCodeBlocks(text);
text = _DoBlockQuotes(text);
text = pluginHooks.postBlockGamut(text, blockGamutHookCallback);
2013-02-05 14:16:51 -05:00
// We already ran _HashHTMLBlocks() before, in Markdown(), but that
// was to escape raw HTML in the original Markdown source. This time,
// we're escaping the markup we've just created, so that we don't wrap
// <p> tags around block-level tags.
text = _HashHTMLBlocks(text);
text = _FormParagraphs(text, doNotUnhash);
return text;
}
function _RunSpanGamut(text) {
//
// These are all the transformations that occur *within* block-level
// tags like paragraphs, headers, and list items.
//
text = pluginHooks.preSpanGamut(text);
2013-02-05 14:16:51 -05:00
text = _DoCodeSpans(text);
text = _EscapeSpecialCharsWithinTagAttributes(text);
text = _EncodeBackslashEscapes(text);
// Process anchor and image tags. Images must come first,
// because ![foo][f] looks like an anchor.
text = _DoImages(text);
text = _DoAnchors(text);
// Make links out of things like `<http://example.com/>`
// Must come after _DoAnchors(), because you can use < and >
// delimiters in inline links like [this](<url>).
text = _DoAutoLinks(text);
text = text.replace(/~P/g, "://"); // put in place to prevent autolinking; reset now
text = _EncodeAmpsAndAngles(text);
text = _DoItalicsAndBold(text);
// Do hard breaks:
text = text.replace(/ +\n/g, " <br>\n");
text = pluginHooks.postSpanGamut(text);
2013-02-05 14:16:51 -05:00
return text;
}
function _EscapeSpecialCharsWithinTagAttributes(text) {
//
// Within tags -- meaning between < and > -- encode [\ ` * _] so they
// don't conflict with their use in Markdown for code, italics and strong.
//
// Build a regex to find HTML tags and comments. See Friedl's
// "Mastering Regular Expressions", 2nd Ed., pp. 200-201.
// SE: changed the comment part of the regex
var regex = /(<[a-z\/!$]("[^"]*"|'[^']*'|[^'">])*>|<!(--(?:|(?:[^>-]|-[^>])(?:[^-]|-[^-])*)--)>)/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) {
//
// 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: "[*+-]" };
function _ProcessListItems(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 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));
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;
}
function _DoCodeBlocks(text) {
//
// Process Markdown `<pre><code>` blocks.
//
/*
text = text.replace(/
(?:\n\n|^)
( // $1 = the code block -- one or more lines, starting with a space/tab
(?:
(?:[ ]{4}|\t) // Lines must start with a tab or a tab-width of spaces - attacklab: g_tab_width
.*\n+
)+
)
(\n*[ ]{0,3}[^ \t\n]|(?=~0)) // attacklab: g_tab_width
/g ,function(){...});
*/
// attacklab: sentinel workarounds for lack of \A and \Z, safari\khtml bug
text += "~0";
text = text.replace(/(?:\n\n|^)((?:(?:[ ]{4}|\t).*\n+)+)(\n*[ ]{0,3}[^ \t\n]|(?=~0))/g,
function (wholeMatch, m1, m2) {
var codeblock = m1;
var nextChar = m2;
codeblock = _EncodeCode(_Outdent(codeblock));
codeblock = _Detab(codeblock);
codeblock = codeblock.replace(/^\n+/g, ""); // trim leading newlines
codeblock = codeblock.replace(/\n+$/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;");
2013-02-05 14:16:51 -05:00
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;
}
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 "";
});
}
return "<" + protocol + link + ">" + tail;
}
2013-02-05 14:16:51 -05:00
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(/(="|<)?\b(https?|ftp)(:\/\/[-A-Z0-9+&@#\/%?=~_|\[\]\(\)!:,\.;]*[-A-Z0-9+&@#\/%=~_|\[\])])(?=$|\W)/gi, handleTrailingParens);
2013-02-05 14:16:51 -05:00
// autolink anything like <http://example.com>
var replacer = function (wholematch, m1) {
m1encoded = m1.replace(/\_\_/, '%5F%5F');
return "<a href=\"" + m1encoded + "\">" + pluginHooks.plainLinkText(m1) + "</a>";
2013-02-05 14:16:51 -05:00
}
2013-02-05 14:16:51 -05:00
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) );
}
);
*/
2013-02-05 14:16:51 -05:00
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
})();