34 |
35 |
36 |
--------------------------------------------------------------------------------
/js/Markdown.Converter.js:
--------------------------------------------------------------------------------
1 | "use strict";
2 | var Markdown;
3 |
4 | if (typeof exports === "object" && typeof require === "function") // we're in a CommonJS (e.g. Node.js) module
5 | Markdown = exports;
6 | else
7 | Markdown = {};
8 |
9 | // The following text is included for historical reasons, but should
10 | // be taken with a pinch of salt; it's not all true anymore.
11 |
12 | //
13 | // Wherever possible, Showdown is a straight, line-by-line port
14 | // of the Perl version of Markdown.
15 | //
16 | // This is not a normal parser design; it's basically just a
17 | // series of string substitutions. It's hard to read and
18 | // maintain this way, but keeping Showdown close to the original
19 | // design makes it easier to port new features.
20 | //
21 | // More importantly, Showdown behaves like markdown.pl in most
22 | // edge cases. So web applications can do client-side preview
23 | // in Javascript, and then build identical HTML on the server.
24 | //
25 | // This port needs the new RegExp functionality of ECMA 262,
26 | // 3rd Edition (i.e. Javascript 1.5). Most modern web browsers
27 | // should do fine. Even with the new regular expression features,
28 | // We do a lot of work to emulate Perl's regex functionality.
29 | // The tricky changes in this file mostly have the "attacklab:"
30 | // label. Major or self-explanatory changes don't.
31 | //
32 | // Smart diff tools like Araxis Merge will be able to match up
33 | // this file with markdown.pl in a useful way. A little tweaking
34 | // helps: in a copy of markdown.pl, replace "#" with "//" and
35 | // replace "$text" with "text". Be sure to ignore whitespace
36 | // and line endings.
37 | //
38 |
39 |
40 | //
41 | // Usage:
42 | //
43 | // var text = "Markdown *rocks*.";
44 | //
45 | // var converter = new Markdown.Converter();
46 | // var html = converter.makeHtml(text);
47 | //
48 | // alert(html);
49 | //
50 | // Note: move the sample code to the bottom of this
51 | // file before uncommenting it.
52 | //
53 |
54 | (function () {
55 |
56 | function identity(x) { return x; }
57 | function returnFalse(x) { return false; }
58 |
59 | function HookCollection() { }
60 |
61 | HookCollection.prototype = {
62 |
63 | chain: function (hookname, func) {
64 | var original = this[hookname];
65 | if (!original)
66 | throw new Error("unknown hook " + hookname);
67 |
68 | if (original === identity)
69 | this[hookname] = func;
70 | else
71 | this[hookname] = function (text) {
72 | var args = Array.prototype.slice.call(arguments, 0);
73 | args[0] = original.apply(null, args);
74 | return func.apply(null, args);
75 | };
76 | },
77 | set: function (hookname, func) {
78 | if (!this[hookname])
79 | throw new Error("unknown hook " + hookname);
80 | this[hookname] = func;
81 | },
82 | addNoop: function (hookname) {
83 | this[hookname] = identity;
84 | },
85 | addFalse: function (hookname) {
86 | this[hookname] = returnFalse;
87 | }
88 | };
89 |
90 | Markdown.HookCollection = HookCollection;
91 |
92 | // g_urls and g_titles allow arbitrary user-entered strings as keys. This
93 | // caused an exception (and hence stopped the rendering) when the user entered
94 | // e.g. [push] or [__proto__]. Adding a prefix to the actual key prevents this
95 | // (since no builtin property starts with "s_"). See
96 | // http://meta.stackexchange.com/questions/64655/strange-wmd-bug
97 | // (granted, switching from Array() to Object() alone would have left only __proto__
98 | // to be a problem)
99 | function SaveHash() { }
100 | SaveHash.prototype = {
101 | set: function (key, value) {
102 | this["s_" + key] = value;
103 | },
104 | get: function (key) {
105 | return this["s_" + key];
106 | }
107 | };
108 |
109 | Markdown.Converter = function (OPTIONS) {
110 | var pluginHooks = this.hooks = new HookCollection();
111 |
112 | // given a URL that was encountered by itself (without markup), should return the link text that's to be given to this link
113 | pluginHooks.addNoop("plainLinkText");
114 |
115 | // called with the orignal text as given to makeHtml. The result of this plugin hook is the actual markdown source that will be cooked
116 | pluginHooks.addNoop("preConversion");
117 |
118 | // called with the text once all normalizations have been completed (tabs to spaces, line endings, etc.), but before any conversions have
119 | pluginHooks.addNoop("postNormalization");
120 |
121 | // Called with the text before / after creating block elements like code blocks and lists. Note that this is called recursively
122 | // with inner content, e.g. it's called with the full text, and then only with the content of a blockquote. The inner
123 | // call will receive outdented text.
124 | pluginHooks.addNoop("preBlockGamut");
125 | pluginHooks.addNoop("postBlockGamut");
126 |
127 | // called with the text of a single block element before / after the span-level conversions (bold, code spans, etc.) have been made
128 | pluginHooks.addNoop("preSpanGamut");
129 | pluginHooks.addNoop("postSpanGamut");
130 |
131 | // called with the final cooked HTML code. The result of this plugin hook is the actual output of makeHtml
132 | pluginHooks.addNoop("postConversion");
133 |
134 | //
135 | // Private state of the converter instance:
136 | //
137 |
138 | // Global hashes, used by various utility routines
139 | var g_urls;
140 | var g_titles;
141 | var g_html_blocks;
142 |
143 | // Used to track when we're inside an ordered or unordered list
144 | // (see _ProcessListItems() for details):
145 | var g_list_level;
146 |
147 | OPTIONS = OPTIONS || {};
148 | var asciify = identity, deasciify = identity;
149 | if (OPTIONS.nonAsciiLetters) {
150 |
151 | /* In JavaScript regular expressions, \w only denotes [a-zA-Z0-9_].
152 | * That's why there's inconsistent handling e.g. with intra-word bolding
153 | * of Japanese words. That's why we do the following if OPTIONS.nonAsciiLetters
154 | * is true:
155 | *
156 | * Before doing bold and italics, we find every instance
157 | * of a unicode word character in the Markdown source that is not
158 | * matched by \w, and the letter "Q". We take the character's code point
159 | * and encode it in base 51, using the "digits"
160 | *
161 | * A, B, ..., P, R, ..., Y, Z, a, b, ..., y, z
162 | *
163 | * delimiting it with "Q" on both sides. For example, the source
164 | *
165 | * > In Chinese, the smurfs are called 藍精靈, meaning "blue spirits".
166 | *
167 | * turns into
168 | *
169 | * > In Chinese, the smurfs are called QNIhQQMOIQQOuUQ, meaning "blue spirits".
170 | *
171 | * Since everything that is a letter in Unicode is now a letter (or
172 | * several letters) in ASCII, \w and \b should always do the right thing.
173 | *
174 | * After the bold/italic conversion, we decode again; since "Q" was encoded
175 | * alongside all non-ascii characters (as "QBfQ"), and the conversion
176 | * will not generate "Q", the only instances of that letter should be our
177 | * encoded characters. And since the conversion will not break words, the
178 | * "Q...Q" should all still be in one piece.
179 | *
180 | * We're using "Q" as the delimiter because it's probably one of the
181 | * rarest characters, and also because I can't think of any special behavior
182 | * that would ever be triggered by this letter (to use a silly example, if we
183 | * delimited with "H" on the left and "P" on the right, then "Ψ" would be
184 | * encoded as "HTTP", which may cause special behavior). The latter would not
185 | * actually be a huge issue for bold/italic, but may be if we later use it
186 | * in other places as well.
187 | * */
188 | (function () {
189 | var lettersThatJavaScriptDoesNotKnowAndQ = /[Q\u00aa\u00b5\u00ba\u00c0-\u00d6\u00d8-\u00f6\u00f8-\u02c1\u02c6-\u02d1\u02e0-\u02e4\u02ec\u02ee\u0370-\u0374\u0376-\u0377\u037a-\u037d\u0386\u0388-\u038a\u038c\u038e-\u03a1\u03a3-\u03f5\u03f7-\u0481\u048a-\u0523\u0531-\u0556\u0559\u0561-\u0587\u05d0-\u05ea\u05f0-\u05f2\u0621-\u064a\u0660-\u0669\u066e-\u066f\u0671-\u06d3\u06d5\u06e5-\u06e6\u06ee-\u06fc\u06ff\u0710\u0712-\u072f\u074d-\u07a5\u07b1\u07c0-\u07ea\u07f4-\u07f5\u07fa\u0904-\u0939\u093d\u0950\u0958-\u0961\u0966-\u096f\u0971-\u0972\u097b-\u097f\u0985-\u098c\u098f-\u0990\u0993-\u09a8\u09aa-\u09b0\u09b2\u09b6-\u09b9\u09bd\u09ce\u09dc-\u09dd\u09df-\u09e1\u09e6-\u09f1\u0a05-\u0a0a\u0a0f-\u0a10\u0a13-\u0a28\u0a2a-\u0a30\u0a32-\u0a33\u0a35-\u0a36\u0a38-\u0a39\u0a59-\u0a5c\u0a5e\u0a66-\u0a6f\u0a72-\u0a74\u0a85-\u0a8d\u0a8f-\u0a91\u0a93-\u0aa8\u0aaa-\u0ab0\u0ab2-\u0ab3\u0ab5-\u0ab9\u0abd\u0ad0\u0ae0-\u0ae1\u0ae6-\u0aef\u0b05-\u0b0c\u0b0f-\u0b10\u0b13-\u0b28\u0b2a-\u0b30\u0b32-\u0b33\u0b35-\u0b39\u0b3d\u0b5c-\u0b5d\u0b5f-\u0b61\u0b66-\u0b6f\u0b71\u0b83\u0b85-\u0b8a\u0b8e-\u0b90\u0b92-\u0b95\u0b99-\u0b9a\u0b9c\u0b9e-\u0b9f\u0ba3-\u0ba4\u0ba8-\u0baa\u0bae-\u0bb9\u0bd0\u0be6-\u0bef\u0c05-\u0c0c\u0c0e-\u0c10\u0c12-\u0c28\u0c2a-\u0c33\u0c35-\u0c39\u0c3d\u0c58-\u0c59\u0c60-\u0c61\u0c66-\u0c6f\u0c85-\u0c8c\u0c8e-\u0c90\u0c92-\u0ca8\u0caa-\u0cb3\u0cb5-\u0cb9\u0cbd\u0cde\u0ce0-\u0ce1\u0ce6-\u0cef\u0d05-\u0d0c\u0d0e-\u0d10\u0d12-\u0d28\u0d2a-\u0d39\u0d3d\u0d60-\u0d61\u0d66-\u0d6f\u0d7a-\u0d7f\u0d85-\u0d96\u0d9a-\u0db1\u0db3-\u0dbb\u0dbd\u0dc0-\u0dc6\u0e01-\u0e30\u0e32-\u0e33\u0e40-\u0e46\u0e50-\u0e59\u0e81-\u0e82\u0e84\u0e87-\u0e88\u0e8a\u0e8d\u0e94-\u0e97\u0e99-\u0e9f\u0ea1-\u0ea3\u0ea5\u0ea7\u0eaa-\u0eab\u0ead-\u0eb0\u0eb2-\u0eb3\u0ebd\u0ec0-\u0ec4\u0ec6\u0ed0-\u0ed9\u0edc-\u0edd\u0f00\u0f20-\u0f29\u0f40-\u0f47\u0f49-\u0f6c\u0f88-\u0f8b\u1000-\u102a\u103f-\u1049\u1050-\u1055\u105a-\u105d\u1061\u1065-\u1066\u106e-\u1070\u1075-\u1081\u108e\u1090-\u1099\u10a0-\u10c5\u10d0-\u10fa\u10fc\u1100-\u1159\u115f-\u11a2\u11a8-\u11f9\u1200-\u1248\u124a-\u124d\u1250-\u1256\u1258\u125a-\u125d\u1260-\u1288\u128a-\u128d\u1290-\u12b0\u12b2-\u12b5\u12b8-\u12be\u12c0\u12c2-\u12c5\u12c8-\u12d6\u12d8-\u1310\u1312-\u1315\u1318-\u135a\u1380-\u138f\u13a0-\u13f4\u1401-\u166c\u166f-\u1676\u1681-\u169a\u16a0-\u16ea\u1700-\u170c\u170e-\u1711\u1720-\u1731\u1740-\u1751\u1760-\u176c\u176e-\u1770\u1780-\u17b3\u17d7\u17dc\u17e0-\u17e9\u1810-\u1819\u1820-\u1877\u1880-\u18a8\u18aa\u1900-\u191c\u1946-\u196d\u1970-\u1974\u1980-\u19a9\u19c1-\u19c7\u19d0-\u19d9\u1a00-\u1a16\u1b05-\u1b33\u1b45-\u1b4b\u1b50-\u1b59\u1b83-\u1ba0\u1bae-\u1bb9\u1c00-\u1c23\u1c40-\u1c49\u1c4d-\u1c7d\u1d00-\u1dbf\u1e00-\u1f15\u1f18-\u1f1d\u1f20-\u1f45\u1f48-\u1f4d\u1f50-\u1f57\u1f59\u1f5b\u1f5d\u1f5f-\u1f7d\u1f80-\u1fb4\u1fb6-\u1fbc\u1fbe\u1fc2-\u1fc4\u1fc6-\u1fcc\u1fd0-\u1fd3\u1fd6-\u1fdb\u1fe0-\u1fec\u1ff2-\u1ff4\u1ff6-\u1ffc\u203f-\u2040\u2054\u2071\u207f\u2090-\u2094\u2102\u2107\u210a-\u2113\u2115\u2119-\u211d\u2124\u2126\u2128\u212a-\u212d\u212f-\u2139\u213c-\u213f\u2145-\u2149\u214e\u2183-\u2184\u2c00-\u2c2e\u2c30-\u2c5e\u2c60-\u2c6f\u2c71-\u2c7d\u2c80-\u2ce4\u2d00-\u2d25\u2d30-\u2d65\u2d6f\u2d80-\u2d96\u2da0-\u2da6\u2da8-\u2dae\u2db0-\u2db6\u2db8-\u2dbe\u2dc0-\u2dc6\u2dc8-\u2dce\u2dd0-\u2dd6\u2dd8-\u2dde\u2e2f\u3005-\u3006\u3031-\u3035\u303b-\u303c\u3041-\u3096\u309d-\u309f\u30a1-\u30fa\u30fc-\u30ff\u3105-\u312d\u3131-\u318e\u31a0-\u31b7\u31f0-\u31ff\u3400-\u4db5\u4e00-\u9fc3\ua000-\ua48c\ua500-\ua60c\ua610-\ua62b\ua640-\ua65f\ua662-\ua66e\ua67f-\ua697\ua717-\ua71f\ua722-\ua788\ua78b-\ua78c\ua7fb-\ua801\ua803-\ua805\ua807-\ua80a\ua80c-\ua822\ua840-\ua873\ua882-\ua8b3\ua8d0-\ua8d9\ua900-\ua925\ua930-\ua946\uaa00-\uaa28\uaa40-\uaa42\uaa44-\uaa4b\uaa50-\uaa59\uac00-\ud7a3\uf900-\ufa2d\ufa30-\ufa6a\ufa70-\ufad9\ufb00-\ufb06\ufb13-\ufb17\ufb1d\ufb1f-\ufb28\ufb2a-\ufb36\ufb38-\ufb3c\ufb3e\ufb40-\ufb41\ufb43-\ufb44\ufb46-\ufbb1\ufbd3-\ufd3d\ufd50-\ufd8f\ufd92-\ufdc7\ufdf0-\ufdfb\ufe33-\ufe34\ufe4d-\ufe4f\ufe70-\ufe74\ufe76-\ufefc\uff10-\uff19\uff21-\uff3a\uff3f\uff41-\uff5a\uff66-\uffbe\uffc2-\uffc7\uffca-\uffcf\uffd2-\uffd7\uffda-\uffdc]/g;
190 | var cp_Q = "Q".charCodeAt(0);
191 | var cp_A = "A".charCodeAt(0);
192 | var cp_Z = "Z".charCodeAt(0);
193 | var dist_Za = "a".charCodeAt(0) - cp_Z - 1;
194 |
195 | asciify = function(text) {
196 | return text.replace(lettersThatJavaScriptDoesNotKnowAndQ, function (m) {
197 | var c = m.charCodeAt(0);
198 | var s = "";
199 | var v;
200 | while (c > 0) {
201 | v = (c % 51) + cp_A;
202 | if (v >= cp_Q)
203 | v++;
204 | if (v > cp_Z)
205 | v += dist_Za;
206 | s = String.fromCharCode(v) + s;
207 | c = c / 51 | 0;
208 | }
209 | return "Q" + s + "Q";
210 | })
211 | };
212 |
213 | deasciify = function(text) {
214 | return text.replace(/Q([A-PR-Za-z]{1,3})Q/g, function (m, s) {
215 | var c = 0;
216 | var v;
217 | for (var i = 0; i < s.length; i++) {
218 | v = s.charCodeAt(i);
219 | if (v > cp_Z)
220 | v -= dist_Za;
221 | if (v > cp_Q)
222 | v--;
223 | v -= cp_A;
224 | c = (c * 51) + v;
225 | }
226 | return String.fromCharCode(c);
227 | })
228 | }
229 | })();
230 | }
231 |
232 | this.makeHtml = function (text) {
233 |
234 | //
235 | // Main function. The order in which other subs are called here is
236 | // essential. Link and image substitutions need to happen before
237 | // _EscapeSpecialCharsWithinTagAttributes(), so that any *'s or _'s in the
238 | // and s around 362 | // "paragraphs" that are wrapped in non-block-level tags, such as anchors, 363 | // phrase emphasis, and spans. The list of tags we're looking for is 364 | // hard-coded: 365 | var block_tags_a = "p|div|h[1-6]|blockquote|pre|table|dl|ol|ul|script|noscript|form|fieldset|iframe|math|ins|del" 366 | var block_tags_b = "p|div|h[1-6]|blockquote|pre|table|dl|ol|ul|script|noscript|form|fieldset|iframe|math" 367 | 368 | // First, look for nested blocks, e.g.: 369 | //
tags around block-level tags.
516 | text = _HashHTMLBlocks(text);
517 | text = _FormParagraphs(text, doNotUnhash);
518 |
519 | return text;
520 | }
521 |
522 | function _RunSpanGamut(text) {
523 | //
524 | // These are all the transformations that occur *within* block-level
525 | // tags like paragraphs, headers, and list items.
526 | //
527 |
528 | text = pluginHooks.preSpanGamut(text);
529 |
530 | text = _DoCodeSpans(text);
531 | text = _EscapeSpecialCharsWithinTagAttributes(text);
532 | text = _EncodeBackslashEscapes(text);
533 |
534 | // Process anchor and image tags. Images must come first,
535 | // because ![foo][f] looks like an anchor.
536 | text = _DoImages(text);
537 | text = _DoAnchors(text);
538 |
539 | // Make links out of things like ` Just type tags
1253 | //
1254 |
1255 | // Strip leading and trailing lines:
1256 | text = text.replace(/^\n+/g, "");
1257 | text = text.replace(/\n+$/g, "");
1258 |
1259 | var grafs = text.split(/\n{2,}/g);
1260 | var grafsOut = [];
1261 |
1262 | var markerRe = /~K(\d+)K/;
1263 |
1264 | //
1265 | // Wrap tags.
1266 | //
1267 | var end = grafs.length;
1268 | for (var i = 0; i < end; i++) {
1269 | var str = grafs[i];
1270 |
1271 | // if this is an HTML marker, copy it
1272 | if (markerRe.test(str)) {
1273 | grafsOut.push(str);
1274 | }
1275 | else if (/\S/.test(str)) {
1276 | str = _RunSpanGamut(str);
1277 | str = str.replace(/^([ \t]*)/g, " ");
1278 | str += "
\n");
551 |
552 | text = pluginHooks.postSpanGamut(text);
553 |
554 | return text;
555 | }
556 |
557 | function _EscapeSpecialCharsWithinTagAttributes(text) {
558 | //
559 | // Within tags -- meaning between < and > -- encode [\ ` * _] so they
560 | // don't conflict with their use in Markdown for code, italics and strong.
561 | //
562 |
563 | // Build a regex to find HTML tags and comments. See Friedl's
564 | // "Mastering Regular Expressions", 2nd Ed., pp. 200-201.
565 |
566 | // SE: changed the comment part of the regex
567 |
568 | var regex = /(<[a-z\/!$]("[^"]*"|'[^']*'|[^'">])*>|-]|-[^>])(?:[^-]|-[^-])*)--)>)/gi;
569 |
570 | text = text.replace(regex, function (wholeMatch) {
571 | var tag = wholeMatch.replace(/(.)<\/?code>(?=.)/g, "$1`");
572 | tag = escapeCharacters(tag, wholeMatch.charAt(1) == "!" ? "\\`*_/" : "\\`*_"); // also escape slashes in comments to prevent autolinking there -- http://meta.stackexchange.com/questions/95987
573 | return tag;
574 | });
575 |
576 | return text;
577 | }
578 |
579 | function _DoAnchors(text) {
580 | //
581 | // Turn Markdown link shortcuts into XHTML tags.
582 | //
583 | //
584 | // First, handle reference-style links: [link text] [id]
585 | //
586 |
587 | /*
588 | text = text.replace(/
589 | ( // wrap whole match in $1
590 | \[
591 | (
592 | (?:
593 | \[[^\]]*\] // allow brackets nested one level
594 | |
595 | [^\[] // or anything else
596 | )*
597 | )
598 | \]
599 |
600 | [ ]? // one optional space
601 | (?:\n[ ]*)? // one optional newline followed by spaces
602 |
603 | \[
604 | (.*?) // id = $3
605 | \]
606 | )
607 | ()()()() // pad remaining backreferences
608 | /g, writeAnchorTag);
609 | */
610 | text = text.replace(/(\[((?:\[[^\]]*\]|[^\[\]])*)\][ ]?(?:\n[ ]*)?\[(.*?)\])()()()()/g, writeAnchorTag);
611 |
612 | //
613 | // Next, inline-style links: [link text](url "optional title")
614 | //
615 |
616 | /*
617 | text = text.replace(/
618 | ( // wrap whole match in $1
619 | \[
620 | (
621 | (?:
622 | \[[^\]]*\] // allow brackets nested one level
623 | |
624 | [^\[\]] // or anything else
625 | )*
626 | )
627 | \]
628 | \( // literal paren
629 | [ \t]*
630 | () // no id, so leave $3 empty
631 | ?
638 | [ \t]*
639 | ( // $5
640 | (['"]) // quote char = $6
641 | (.*?) // Title = $7
642 | \6 // matching quote
643 | [ \t]* // ignore any spaces/tabs between closing quote and )
644 | )? // title is optional
645 | \)
646 | )
647 | /g, writeAnchorTag);
648 | */
649 |
650 | text = text.replace(/(\[((?:\[[^\]]*\]|[^\[\]])*)\]\([ \t]*()?[ \t]*((['"])(.*?)\6[ \t]*)?\))/g, writeAnchorTag);
651 |
652 | //
653 | // Last, handle reference-style shortcuts: [link text]
654 | // These must come last in case you've also got [link test][1]
655 | // or [link test](/foo)
656 | //
657 |
658 | /*
659 | text = text.replace(/
660 | ( // wrap whole match in $1
661 | \[
662 | ([^\[\]]+) // link text = $2; can't contain '[' or ']'
663 | \]
664 | )
665 | ()()()()() // pad rest of backreferences
666 | /g, writeAnchorTag);
667 | */
668 | text = text.replace(/(\[([^\[\]]+)\])()()()()()/g, writeAnchorTag);
669 |
670 | return text;
671 | }
672 |
673 | function writeAnchorTag(wholeMatch, m1, m2, m3, m4, m5, m6, m7) {
674 | if (m7 == undefined) m7 = "";
675 | var whole_match = m1;
676 | var link_text = m2.replace(/:\/\//g, "~P"); // to prevent auto-linking withing the link. will be converted back after the auto-linker runs
677 | var link_id = m3.toLowerCase();
678 | var url = m4;
679 | var title = m7;
680 |
681 | if (url == "") {
682 | if (link_id == "") {
683 | // lower-case and turn embedded newlines into spaces
684 | link_id = link_text.toLowerCase().replace(/ ?\n/g, " ");
685 | }
686 | url = "#" + link_id;
687 |
688 | if (g_urls.get(link_id) != undefined) {
689 | url = g_urls.get(link_id);
690 | if (g_titles.get(link_id) != undefined) {
691 | title = g_titles.get(link_id);
692 | }
693 | }
694 | else {
695 | if (whole_match.search(/\(\s*\)$/m) > -1) {
696 | // Special case for explicit empty url
697 | url = "";
698 | } else {
699 | return whole_match;
700 | }
701 | }
702 | }
703 | url = encodeProblemUrlChars(url);
704 | url = escapeCharacters(url, "*_");
705 | var result = "";
714 |
715 | return result;
716 | }
717 |
718 | function _DoImages(text) {
719 | //
720 | // Turn Markdown image shortcuts into tags.
721 | //
722 |
723 | //
724 | // First, handle reference-style labeled images: ![alt text][id]
725 | //
726 |
727 | /*
728 | text = text.replace(/
729 | ( // wrap whole match in $1
730 | !\[
731 | (.*?) // alt text = $2
732 | \]
733 |
734 | [ ]? // one optional space
735 | (?:\n[ ]*)? // one optional newline followed by spaces
736 |
737 | \[
738 | (.*?) // id = $3
739 | \]
740 | )
741 | ()()()() // pad rest of backreferences
742 | /g, writeImageTag);
743 | */
744 | text = text.replace(/(!\[(.*?)\][ ]?(?:\n[ ]*)?\[(.*?)\])()()()()/g, writeImageTag);
745 |
746 | //
747 | // Next, handle inline images: 
748 | // Don't forget: encode * and _
749 |
750 | /*
751 | text = text.replace(/
752 | ( // wrap whole match in $1
753 | !\[
754 | (.*?) // alt text = $2
755 | \]
756 | \s? // One optional whitespace character
757 | \( // literal paren
758 | [ \t]*
759 | () // no id, so leave $3 empty
760 | ? // src url = $4
761 | [ \t]*
762 | ( // $5
763 | (['"]) // quote char = $6
764 | (.*?) // title = $7
765 | \6 // matching quote
766 | [ \t]*
767 | )? // title is optional
768 | \)
769 | )
770 | /g, writeImageTag);
771 | */
772 | text = text.replace(/(!\[(.*?)\]\s?\([ \t]*()?[ \t]*((['"])(.*?)\6[ \t]*)?\))/g, writeImageTag);
773 |
774 | return text;
775 | }
776 |
777 | function attributeEncode(text) {
778 | // unconditionally replace angle brackets here -- what ends up in an attribute (e.g. alt or title)
779 | // never makes sense to have verbatim HTML in it (and the sanitizer would totally break it)
780 | return text.replace(/>/g, ">").replace(/" + _RunSpanGamut(m1) + "\n\n"; }
839 | );
840 |
841 | text = text.replace(/^(.+)[ \t]*\n-+[ \t]*\n+/gm,
842 | function (matchFound, m1) { return "
" + _RunSpanGamut(m1) + "
\n\n"; }
843 | );
844 |
845 | // atx-style headers:
846 | // # Header 1
847 | // ## Header 2
848 | // ## Header 2 with closing hashes ##
849 | // ...
850 | // ###### Header 6
851 | //
852 |
853 | /*
854 | text = text.replace(/
855 | ^(\#{1,6}) // $1 = string of #'s
856 | [ \t]*
857 | (.+?) // $2 = Header text
858 | [ \t]*
859 | \#* // optional closing #'s (not counted)
860 | \n+
861 | /gm, function() {...});
862 | */
863 |
864 | text = text.replace(/^(\#{1,6})[ \t]*(.+?)[ \t]*\#*\n+/gm,
865 | function (wholeMatch, m1, m2) {
866 | var h_level = m1.length;
867 | return "` blocks.
1041 | //
1042 |
1043 | /*
1044 | text = text.replace(/
1045 | (?:\n\n|^)
1046 | ( // $1 = the code block -- one or more lines, starting with a space/tab
1047 | (?:
1048 | (?:[ ]{4}|\t) // Lines must start with a tab or a tab-width of spaces - attacklab: g_tab_width
1049 | .*\n+
1050 | )+
1051 | )
1052 | (\n*[ ]{0,3}[^ \t\n]|(?=~0)) // attacklab: g_tab_width
1053 | /g ,function(){...});
1054 | */
1055 |
1056 | // attacklab: sentinel workarounds for lack of \A and \Z, safari\khtml bug
1057 | text += "~0";
1058 |
1059 | text = text.replace(/(?:\n\n|^\n?)((?:(?:[ ]{4}|\t).*\n+)+)(\n*[ ]{0,3}[^ \t\n]|(?=~0))/g,
1060 | function (wholeMatch, m1, m2) {
1061 | var codeblock = m1;
1062 | var nextChar = m2;
1063 |
1064 | codeblock = _EncodeCode(_Outdent(codeblock));
1065 | codeblock = _Detab(codeblock);
1066 | codeblock = codeblock.replace(/^\n+/g, ""); // trim leading newlines
1067 | codeblock = codeblock.replace(/\n+$/g, ""); // trim trailing whitespace
1068 |
1069 | codeblock = "
";
1070 |
1071 | return "\n\n" + codeblock + "\n\n" + nextChar;
1072 | }
1073 | );
1074 |
1075 | // attacklab: strip sentinel
1076 | text = text.replace(/~0/, "");
1077 |
1078 | return text;
1079 | }
1080 |
1081 | function _DoCodeSpans(text) {
1082 | //
1083 | // * Backtick quotes are used for " + codeblock + "\n spans.
1084 | //
1085 | // * You can use multiple backticks as the delimiters if you want to
1086 | // include literal backticks in the code span. So, this input:
1087 | //
1088 | // Just type ``foo `bar` baz`` at the prompt.
1089 | //
1090 | // Will translate to:
1091 | //
1092 | // foo `bar` baz at the prompt.`bar` ...
1105 | //
1106 |
1107 | /*
1108 | text = text.replace(/
1109 | (^|[^\\`]) // Character before opening ` can't be a backslash or backtick
1110 | (`+) // $2 = Opening run of `
1111 | (?!`) // and no more backticks -- match the full run
1112 | ( // $3 = The code block
1113 | [^\r]*?
1114 | [^`] // attacklab: work around lack of lookbehind
1115 | )
1116 | \2 // Matching closer
1117 | (?!`)
1118 | /gm, function(){...});
1119 | */
1120 |
1121 | text = text.replace(/(^|[^\\`])(`+)(?!`)([^\r]*?[^`])\2(?!`)/gm,
1122 | function (wholeMatch, m1, m2, m3, m4) {
1123 | var c = m3;
1124 | c = c.replace(/^([ \t]*)/g, ""); // leading whitespace
1125 | c = c.replace(/[ \t]*$/g, ""); // trailing whitespace
1126 | c = _EncodeCode(c);
1127 | 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.
1128 | return m1 + "" + c + "";
1129 | }
1130 | );
1131 |
1132 | return text;
1133 | }
1134 |
1135 | function _EncodeCode(text) {
1136 | //
1137 | // Encode/escape certain characters inside Markdown code runs.
1138 | // The point is that in code, these characters are literals,
1139 | // and lose their special Markdown meanings.
1140 | //
1141 | // Encode all ampersands; HTML entities are not
1142 | // entities within a Markdown code span.
1143 | text = text.replace(/&/g, "&");
1144 |
1145 | // Do the angle bracket song and dance:
1146 | text = text.replace(//g, ">");
1148 |
1149 | // Now, escape characters that are magic in Markdown:
1150 | text = escapeCharacters(text, "\*_{}[]\\", false);
1151 |
1152 | // jj the line above breaks this:
1153 | //---
1154 |
1155 | //* Item
1156 |
1157 | // 1. Subitem
1158 |
1159 | // special char: *
1160 | //---
1161 |
1162 | return text;
1163 | }
1164 |
1165 | function _DoItalicsAndBold(text) {
1166 |
1167 | text = asciify(text);
1168 |
1169 | // must go first:
1170 |
1171 | // (^|[\W_]) Start with a non-letter or beginning of string. Store in \1.
1172 | // (?:(?!\1)|(?=^)) Either the next character is *not* the same as the previous,
1173 | // or we started at the end of the string (in which case the previous
1174 | // group had zero width, so we're still there). Because the next
1175 | // character is the marker, this means that if there are e.g. multiple
1176 | // underscores in a row, we can only match the left-most ones (which
1177 | // prevents foo___bar__ from getting bolded)
1178 | // (\*|_) The marker character itself, asterisk or underscore. Store in \2.
1179 | // \2 The marker again, since bold needs two.
1180 | // (?=\S) The first bolded character cannot be a space.
1181 | // ([^\r]*?\S) The actual bolded string. At least one character, and it cannot *end*
1182 | // with a space either. Note that like in many other places, [^\r] is
1183 | // just a workaround for JS' lack of single-line regexes; it's equivalent
1184 | // to a . in an /s regex, because the string cannot contain any \r (they
1185 | // are removed in the normalizing step).
1186 | // \2\2 The marker character, twice -- end of bold.
1187 | // (?!\2) Not followed by another marker character (ensuring that we match the
1188 | // rightmost two in a longer row)...
1189 | // (?=[\W_]|$) ...but by any other non-word character or the end of string.
1190 | text = text.replace(/(^|[\W_])(?:(?!\1)|(?=^))(\*|_)\2(?=\S)([^\r]*?\S)\2\2(?!\2)(?=[\W_]|$)/g,
1191 | "$1$3");
1192 |
1193 | // This is almost identical to the regex, except 1) there's obviously just one marker
1194 | // character, and 2) the italicized string cannot contain the marker character.
1195 | text = text.replace(/(^|[\W_])(?:(?!\1)|(?=^))(\*|_)(?=\S)((?:(?!\2)[^\r])*?\S)\2(?!\2)(?=[\W_]|$)/g,
1196 | "$1$3");
1197 |
1198 | return deasciify(text);
1199 | }
1200 |
1201 | function _DoBlockQuotes(text) {
1202 |
1203 | /*
1204 | text = text.replace(/
1205 | ( // Wrap whole match in $1
1206 | (
1207 | ^[ \t]*>[ \t]? // '>' at the start of a line
1208 | .+\n // rest of the first line
1209 | (.+\n)* // subsequent consecutive lines
1210 | \n* // blanks
1211 | )+
1212 | )
1213 | /gm, function(){...});
1214 | */
1215 |
1216 | text = text.replace(/((^[ \t]*>[ \t]?.+\n(.+\n)*\n*)+)/gm,
1217 | function (wholeMatch, m1) {
1218 | var bq = m1;
1219 |
1220 | // attacklab: hack around Konqueror 3.5.4 bug:
1221 | // "----------bug".replace(/^-/g,"") == "bug"
1222 |
1223 | bq = bq.replace(/^[ \t]*>[ \t]?/gm, "~0"); // trim one level of quoting
1224 |
1225 | // attacklab: clean up hack
1226 | bq = bq.replace(/~0/g, "");
1227 |
1228 | bq = bq.replace(/^[ \t]+$/gm, ""); // trim whitespace-only lines
1229 | bq = _RunBlockGamut(bq); // recurse
1230 |
1231 | bq = bq.replace(/(^|\n)/g, "$1 ");
1232 | // These leading spaces screw with content, so we need to fix that:
1233 | bq = bq.replace(
1234 | /(\s*
[^\r]+?<\/pre>)/gm,
1235 | function (wholeMatch, m1) {
1236 | var pre = m1;
1237 | // attacklab: hack around Konqueror 3.5.4 bug:
1238 | pre = pre.replace(/^ /mg, "~0");
1239 | pre = pre.replace(/~0/g, "");
1240 | return pre;
1241 | });
1242 |
1243 | return hashBlock("\n" + bq + "\n
");
1244 | }
1245 | );
1246 | return text;
1247 | }
1248 |
1249 | function _FormParagraphs(text, doNotUnhash) {
1250 | //
1251 | // Params:
1252 | // $text - string to process with html