From 719f6a1165d5852cf22cc4c28ea7b30027bd7198 Mon Sep 17 00:00:00 2001 From: Christian Mueller Date: Sat, 28 Sep 2013 22:21:55 +0200 Subject: [PATCH] new markdown converter version --- .../public/js/pagedown/Markdown.Converter.js | 106 +++++++++++++++--- 1 file changed, 93 insertions(+), 13 deletions(-) diff --git a/resources/public/js/pagedown/Markdown.Converter.js b/resources/public/js/pagedown/Markdown.Converter.js index 58fc54a..f5a0cd2 100644 --- a/resources/public/js/pagedown/Markdown.Converter.js +++ b/resources/public/js/pagedown/Markdown.Converter.js @@ -67,7 +67,11 @@ else if (original === identity) this[hookname] = func; else - this[hookname] = function (x) { return func(original(x)); } + 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]) @@ -103,9 +107,28 @@ else Markdown.Converter = function () { var pluginHooks = this.hooks = new HookCollection(); - pluginHooks.addNoop("plainLinkText"); // 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("preConversion"); // 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("postConversion"); // called with the final cooked HTML code. The result of this plugin hook is the actual output of makeHtml + + // 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: @@ -168,6 +191,8 @@ else // 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); @@ -378,12 +403,17 @@ else 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: @@ -395,6 +425,8 @@ else 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, @@ -412,6 +444,8 @@ else // tags like paragraphs, headers, and list items. // + text = pluginHooks.preSpanGamut(text); + text = _DoCodeSpans(text); text = _EscapeSpecialCharsWithinTagAttributes(text); text = _EncodeBackslashEscapes(text); @@ -433,6 +467,8 @@ else // Do hard breaks: text = text.replace(/ +\n/g, "
\n"); + + text = pluginHooks.postSpanGamut(text); return text; } @@ -754,7 +790,7 @@ else return text; } - function _DoLists(text) { + function _DoLists(text, isInsideParagraphlessListItem) { // // Form HTML ordered (numbered) and unordered (bulleted) lists. // @@ -794,7 +830,7 @@ else var list = m1; var list_type = (m2.search(/[*+-]/g) > -1) ? "ul" : "ol"; - var result = _ProcessListItems(list, list_type); + var result = _ProcessListItems(list, list_type, isInsideParagraphlessListItem); // Trim any trailing whitespace, to put the closing `` // up on the preceding line, to get it past the current stupid @@ -825,7 +861,7 @@ else var _listItemMarkers = { ol: "\\d+[.]", ul: "[*+-]" }; - function _ProcessListItems(list_str, list_type) { + function _ProcessListItems(list_str, list_type, isInsideParagraphlessListItem) { // // Process the contents of a single ordered or unordered list, splitting it // into individual list items. @@ -901,9 +937,10 @@ else } else { // Recursion for sub-lists: - item = _DoLists(_Outdent(item)); + item = _DoLists(_Outdent(item), /* isInsideParagraphlessListItem= */ true); item = item.replace(/\n$/, ""); // chomp(item) - item = _RunSpanGamut(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 "
  • " + item + "
    • \n"; @@ -938,7 +975,7 @@ else // 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, + 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; @@ -1169,7 +1206,7 @@ else text = text.replace(/&(?!#?[xX]?(?:[0-9a-fA-F]+|\w+);)/g, "&"); // Encode naked <'s - text = text.replace(/<(?![a-z\/?\$!])/gi, "<"); + text = text.replace(/<(?![a-z\/?!]|~D)/gi, "<"); return text; } @@ -1195,14 +1232,57 @@ else 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 // *except* for the case // automatically add < and > around unadorned raw hyperlinks - // must be preceded by space/BOF and followed by non-word/EOF character - text = text.replace(/(^|\s)(https?|ftp)(:\/\/[-A-Z0-9+&@#\/%?=~_|\[\]\(\)!:,\.;]*[-A-Z0-9+&@#\/%=~_|\[\]])($|\W)/gi, "$1<$2$3>$4"); + // 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