Stackedit/js/libs/Markdown.Extra.js

610 lines
22 KiB
JavaScript
Raw Normal View History

2013-04-15 16:19:47 +00:00
(function () {
// A quick way to make sure we're only keeping span-level tags when we need to.
// This isn't supposed to be foolproof. It's just a quick way to make sure we
// keep all span-level tags returned by a pagedown converter. It should allow
// all span-level tags through, with or without attributes.
var inlineTags = new RegExp(['^(<\\/?(a|abbr|acronym|applet|area|b|basefont|',
'bdo|big|button|cite|code|del|dfn|em|figcaption|',
'font|i|iframe|img|input|ins|kbd|label|map|',
'mark|meter|object|param|progress|q|ruby|rp|rt|s|',
'samp|script|select|small|span|strike|strong|',
'sub|sup|textarea|time|tt|u|var|wbr)[^>]*>|',
'<(br)\\s?\\/?>)$'].join(''), 'i');
/******************************************************************
* Utility Functions *
*****************************************************************/
// patch for ie7
if (!Array.indexOf) {
2013-04-26 23:08:13 +00:00
Array.prototype.indexOf = function(obj) {
for (var i = 0; i < this.length; i++) {
if (this[i] == obj) {
return i;
}
}
return -1;
2013-06-03 22:19:52 +00:00
};
2013-04-26 23:08:13 +00:00
}
2013-04-15 16:19:47 +00:00
function trim(str) {
return str.replace(/^\s+|\s+$/g, '');
}
function rtrim(str) {
return str.replace(/\s+$/g, '');
}
// Remove one level of indentation from text. Indent is 4 spaces.
function outdent(text) {
2013-06-22 23:48:57 +00:00
return text.replace(new RegExp('^(\\t|[ ]{1,4})', 'gm'), '');
2013-04-15 16:19:47 +00:00
}
function contains(str, substr) {
return str.indexOf(substr) != -1;
}
// Sanitize html, removing tags that aren't in the whitelist
function sanitizeHtml(html, whitelist) {
return html.replace(/<[^>]*>?/gi, function(tag) {
return tag.match(whitelist) ? tag : '';
});
}
// Merge two arrays, keeping only unique elements.
function union(x, y) {
var obj = {};
for (var i = 0; i < x.length; i++)
obj[x[i]] = x[i];
for (i = 0; i < y.length; i++)
obj[y[i]] = y[i];
var res = [];
for (var k in obj) {
if (obj.hasOwnProperty(k))
res.push(obj[k]);
}
return res;
}
// JS regexes don't support \A or \Z, so we add sentinels, as Pagedown
// does. In this case, we add the ascii codes for start of text (STX) and
// end of text (ETX), an idea borrowed from:
// https://github.com/tanakahisateru/js-markdown-extra
function addAnchors(text) {
if(text.charAt(0) != '\x02')
text = '\x02' + text;
if(text.charAt(text.length - 1) != '\x03')
text = text + '\x03';
return text;
}
// Remove STX and ETX sentinels.
function removeAnchors(text) {
if(text.charAt(0) == '\x02')
text = text.substr(1);
if(text.charAt(text.length - 1) == '\x03')
text = text.substr(0, text.length - 1);
return text;
}
// Convert markdown within an element, retaining only span-level tags
2013-06-02 19:35:44 +00:00
function convertSpans(text, extra) {
return sanitizeHtml(convertAll(text, extra), inlineTags);
2013-04-15 16:19:47 +00:00
}
// Convert internal markdown using the stock pagedown converter
2013-06-02 19:35:44 +00:00
function convertAll(text, extra) {
var result = extra.blockGamutHookCallback(text);
2013-06-02 23:14:44 +00:00
// We need to perform these operations since we skip the steps in the converter
result = unescapeSpecialChars(result);
result = result.replace(/~D/g, "$$").replace(/~T/g, "~");
2013-06-02 19:35:44 +00:00
result = extra.previousPostConversion(result);
return result;
2013-04-15 16:19:47 +00:00
}
// Convert escaped special characters to HTML decimal entity codes.
function processEscapes(text) {
// Markdown extra adds two escapable characters, `:` and `|`
// If escaped, we convert them to html entities so our
// regexes don't recognize them. Markdown doesn't support escaping
// the escape character, e.g. `\\`, which make this even simpler.
return text.replace(/\\\|/g, '&#124;').replace(/\\:/g, '&#58;');
}
2013-06-02 23:14:44 +00:00
// Duplicated from PageDown converter
function unescapeSpecialChars(text) {
// Swap back in all the special characters we've hidden.
2013-06-03 22:19:52 +00:00
text = text.replace(/~E(\d+)E/g, function(wholeMatch, m1) {
2013-06-22 23:48:57 +00:00
var charCodeToReplace = parseInt(m1);
return String.fromCharCode(charCodeToReplace);
});
2013-06-02 23:14:44 +00:00
return text;
}
2013-04-15 16:19:47 +00:00
2013-06-03 22:19:52 +00:00
/*****************************************************************************
2013-06-22 23:48:57 +00:00
* Markdown.Extra *
****************************************************************************/
2013-04-15 16:19:47 +00:00
Markdown.Extra = function() {
// For converting internal markdown (in tables for instance).
// This is necessary since these methods are meant to be called as
// preConversion hooks, and the Markdown converter passed to init()
// won't convert any markdown contained in the html tags we return.
this.converter = null;
// Stores html blocks we generate in hooks so that
// they're not destroyed if the user is using a sanitizing converter
this.hashBlocks = [];
// Special attribute blocks for fenced code blocks and headers enabled.
this.attributeBlocks = false;
// Fenced code block options
this.googleCodePrettify = false;
this.highlightJs = false;
// Table options
this.tableClass = '';
this.tabWidth = 4;
};
Markdown.Extra.init = function(converter, options) {
// Each call to init creates a new instance of Markdown.Extra so it's
// safe to have multiple converters, with different options, on a single page
var extra = new Markdown.Extra();
2013-05-10 10:50:23 +00:00
var postNormalizationTransformations = [];
var preBlockGamutTransformations = [];
2013-06-03 22:19:52 +00:00
var postConversionTransformations = ["unHashExtraBlocks"];
2013-04-15 16:19:47 +00:00
options = options || {};
options.extensions = options.extensions || ["all"];
2013-06-03 22:19:52 +00:00
if (contains(options.extensions, "all")) {
2013-06-22 23:48:57 +00:00
options.extensions = ["tables", "fenced_code_gfm", "def_list", "attr_list"];
2013-06-03 22:19:52 +00:00
}
if (contains(options.extensions, "attr_list")) {
2013-06-22 23:48:57 +00:00
postNormalizationTransformations.push("hashFcbAttributeBlocks");
preBlockGamutTransformations.push("hashHeaderAttributeBlocks");
postConversionTransformations.push("applyAttributeBlocks");
2013-06-03 22:19:52 +00:00
extra.attributeBlocks = true;
}
if (contains(options.extensions, "tables")) {
2013-06-22 23:48:57 +00:00
preBlockGamutTransformations.push("tables");
2013-06-03 22:19:52 +00:00
}
if (contains(options.extensions, "fenced_code_gfm")) {
2013-06-22 23:48:57 +00:00
postNormalizationTransformations.push("fencedCodeBlocks");
2013-06-03 22:19:52 +00:00
}
if (contains(options.extensions, "def_list")) {
preBlockGamutTransformations.push("definitionLists");
}
2013-05-10 10:50:23 +00:00
converter.hooks.chain("postNormalization", function(text) {
2013-06-03 22:19:52 +00:00
return extra.doTransform(postNormalizationTransformations, text) + '\n';
2013-05-10 10:50:23 +00:00
});
2013-04-15 16:19:47 +00:00
2013-06-02 19:35:44 +00:00
converter.hooks.chain("preBlockGamut", function(text, blockGamutHookCallback) {
2013-06-22 23:48:57 +00:00
// Keep a reference to the block gamut callback to run recursively
2013-06-02 19:35:44 +00:00
extra.blockGamutHookCallback = blockGamutHookCallback;
2013-06-03 22:19:52 +00:00
text = processEscapes(text);
return extra.doTransform(preBlockGamutTransformations, text) + '\n';
2013-04-15 16:19:47 +00:00
});
2013-06-03 22:19:52 +00:00
// Keep a reference to the hook chain running before doPostConversion to apply on hashed extra blocks
2013-06-22 23:48:57 +00:00
extra.previousPostConversion = converter.hooks.postConversion;
2013-04-15 16:19:47 +00:00
converter.hooks.chain("postConversion", function(text) {
2013-06-03 22:19:52 +00:00
text = extra.doTransform(postConversionTransformations, text);
// Clear state vars that may use unnecessary memory
this.hashBlocks = [];
return text;
2013-04-15 16:19:47 +00:00
});
if ("highlighter" in options) {
extra.googleCodePrettify = options.highlighter === 'prettify';
extra.highlightJs = options.highlighter === 'highlight';
}
if ("table_class" in options) {
extra.tableClass = options.table_class;
}
2013-06-02 19:35:44 +00:00
extra.converter = converter;
2013-04-15 16:19:47 +00:00
// Caller usually won't need this, but it's handy for testing.
return extra;
};
2013-05-10 10:50:23 +00:00
// Do transformations
Markdown.Extra.prototype.doTransform = function(transformations, text) {
2013-06-22 23:48:57 +00:00
for(var i = 0; i < transformations.length; i++)
text = this[transformations[i]](text);
return text;
2013-04-15 16:19:47 +00:00
};
// Return a placeholder containing a key, which is the block's index in the
// hashBlocks array. We wrap our output in a <p> tag here so Pagedown won't.
Markdown.Extra.prototype.hashExtraBlock = function(block) {
return '\n<p>~X' + (this.hashBlocks.push(block) - 1) + 'X</p>\n';
};
// Replace placeholder blocks in `text` with their corresponding
// html blocks in the hashBlocks array.
Markdown.Extra.prototype.unHashExtraBlocks = function(text) {
var self = this;
2013-06-02 19:35:44 +00:00
function recursiveUnHash() {
var hasHash = false;
text = text.replace(/<p>~X(\d+)X<\/p>/g, function(wholeMatch, m1) {
hasHash = true;
var key = parseInt(m1, 10);
2013-06-02 23:14:44 +00:00
return self.hashBlocks[key];
2013-06-02 19:35:44 +00:00
});
if(hasHash === true) {
recursiveUnHash();
}
}
recursiveUnHash();
2013-04-15 16:19:47 +00:00
return text;
};
/******************************************************************
* Attribute Blocks *
*****************************************************************/
2013-05-10 10:50:23 +00:00
// Extract headers attribute blocks, move them above the element they will be
2013-04-15 16:19:47 +00:00
// applied to, and hash them for later.
2013-05-10 10:50:23 +00:00
Markdown.Extra.prototype.hashHeaderAttributeBlocks = function(text) {
2013-06-22 23:48:57 +00:00
// TODO: use sentinels. Should we just add/remove them in doConversion?
// TODO: better matches for id / class attributes
var attrBlock = "\\{\\s*[.|#][^}]+\\}";
var hdrAttributesA = new RegExp("^(#{1,6}.*#{0,6})\\s+(" + attrBlock + ")[ \\t]*(\\n|0x03)", "gm");
var hdrAttributesB = new RegExp("^(.*)\\s+(" + attrBlock + ")[ \\t]*\\n" +
"(?=[\\-|=]+\\s*(\\n|0x03))", "gm"); // underline lookahead
var self = this;
function attributeCallback(wholeMatch, pre, attr) {
return '<p>~XX' + (self.hashBlocks.push(attr) - 1) + 'XX</p>\n' + pre + "\n";
}
text = text.replace(hdrAttributesA, attributeCallback); // ## headers
text = text.replace(hdrAttributesB, attributeCallback); // underline headers
return text;
2013-05-10 10:50:23 +00:00
};
// Extract FCB attribute blocks, move them above the element they will be
// applied to, and hash them for later.
Markdown.Extra.prototype.hashFcbAttributeBlocks = function(text) {
2013-04-15 16:19:47 +00:00
// TODO: use sentinels. Should we just add/remove them in doConversion?
// TODO: better matches for id / class attributes
2013-06-22 23:48:57 +00:00
var attrBlock = "\\{\\s*[.|#][^}]+\\}";
2013-04-26 23:08:13 +00:00
var fcbAttributes = new RegExp("^(```[^{]*)\\s+(" + attrBlock + ")[ \\t]*\\n" +
2013-04-15 16:19:47 +00:00
"(?=([\\s\\S]*?)\\n```\\s*(\\n|0x03))", "gm");
2013-06-22 23:48:57 +00:00
var self = this;
function attributeCallback(wholeMatch, pre, attr) {
return '<p>~XX' + (self.hashBlocks.push(attr) - 1) + 'XX</p>\n' + pre + "\n";
}
2013-04-15 16:19:47 +00:00
return text.replace(fcbAttributes, attributeCallback);
};
Markdown.Extra.prototype.applyAttributeBlocks = function(text) {
var self = this;
2013-04-21 22:28:46 +00:00
var blockRe = new RegExp('<p>~XX(\\d+)XX</p>[\\s]*' +
2013-04-26 23:08:13 +00:00
'(?:<(h[1-6]|pre)(?: +class="(\\S+)")?(>[\\s\\S]*?</\\2>))', "gm");
2013-04-15 16:19:47 +00:00
text = text.replace(blockRe, function(wholeMatch, k, tag, cls, rest) {
if (!tag) // no following header or fenced code block.
return '';
// get attributes list from hash
var key = parseInt(k, 10);
var attributes = self.hashBlocks[key];
// get id
var id = attributes.match(/#[^\s{}]+/g) || [];
var idStr = id[0] ? ' id="' + id[0].substr(1, id[0].length - 1) + '"' : '';
// get classes and merge with existing classes
var classes = attributes.match(/\.[^\s{}]+/g) || [];
for (var i = 0; i < classes.length; i++) // Remove leading dot
classes[i] = classes[i].substr(1, classes[i].length - 1);
var classStr = '';
if (cls)
classes = union(classes, [cls]);
if (classes.length > 0)
classStr = ' class="' + classes.join(' ') + '"';
return "<" + tag + idStr + classStr + rest;
});
return text;
};
/******************************************************************
* Tables *
*****************************************************************/
// Find and convert Markdown Extra tables into html.
Markdown.Extra.prototype.tables = function(text) {
var self = this;
var leadingPipe = new RegExp(
['^' ,
'[ ]{0,3}' , // Allowed whitespace
'[|]' , // Initial pipe
'(.+)\\n' , // $1: Header Row
'[ ]{0,3}' , // Allowed whitespace
'[|]([ ]*[-:]+[-| :]*)\\n' , // $2: Separator
'(' , // $3: Table Body
'(?:[ ]*[|].*\\n?)*' , // Table rows
')',
'(?:\\n|$)' // Stop at final newline
].join(''),
'gm'
);
var noLeadingPipe = new RegExp(
['^' ,
'[ ]{0,3}' , // Allowed whitespace
'(\\S.*[|].*)\\n' , // $1: Header Row
'[ ]{0,3}' , // Allowed whitespace
'([-:]+[ ]*[|][-| :]*)\\n' , // $2: Separator
'(' , // $3: Table Body
'(?:.*[|].*\\n?)*' , // Table rows
')' ,
'(?:\\n|$)' // Stop at final newline
].join(''),
'gm'
);
text = text.replace(leadingPipe, doTable);
text = text.replace(noLeadingPipe, doTable);
// $1 = header, $2 = separator, $3 = body
function doTable(match, header, separator, body, offset, string) {
// remove any leading pipes and whitespace
header = header.replace(/^ *[|]/m, '');
separator = separator.replace(/^ *[|]/m, '');
body = body.replace(/^ *[|]/gm, '');
// remove trailing pipes and whitespace
header = header.replace(/[|] *$/m, '');
separator = separator.replace(/[|] *$/m, '');
body = body.replace(/[|] *$/gm, '');
// determine column alignments
alignspecs = separator.split(/ *[|] */);
align = [];
for (var i = 0; i < alignspecs.length; i++) {
var spec = alignspecs[i];
if (spec.match(/^ *-+: *$/m))
align[i] = ' style="text-align:right;"';
else if (spec.match(/^ *:-+: *$/m))
align[i] = ' style="text-align:center;"';
else if (spec.match(/^ *:-+ *$/m))
align[i] = ' style="text-align:left;"';
else align[i] = '';
}
// TODO: parse spans in header and rows before splitting, so that pipes
// inside of tags are not interpreted as separators
var headers = header.split(/ *[|] */);
var colCount = headers.length;
// build html
var cls = self.tableClass ? ' class="' + self.tableClass + '"' : '';
var html = ['<table', cls, '>\n', '<thead>\n', '<tr>\n'].join('');
// build column headers.
for (i = 0; i < colCount; i++) {
2013-06-02 19:35:44 +00:00
var headerHtml = convertSpans(trim(headers[i]), self);
2013-04-15 16:19:47 +00:00
html += [" <th", align[i], ">", headerHtml, "</th>\n"].join('');
}
html += "</tr>\n</thead>\n";
// build rows
var rows = body.split('\n');
for (i = 0; i < rows.length; i++) {
if (rows[i].match(/^\s*$/)) // can apply to final row
continue;
// ensure number of rowCells matches colCount
var rowCells = rows[i].split(/ *[|] */);
var lenDiff = colCount - rowCells.length;
for (var j = 0; j < lenDiff; j++)
rowCells.push('');
html += "<tr>\n";
for (j = 0; j < colCount; j++) {
2013-06-02 19:35:44 +00:00
var colHtml = convertSpans(trim(rowCells[j]), self);
2013-04-15 16:19:47 +00:00
html += [" <td", align[j], ">", colHtml, "</td>\n"].join('');
}
html += "</tr>\n";
}
html += "</table>\n";
// replace html with placeholder until postConversion step
return self.hashExtraBlock(html);
}
return text;
};
/******************************************************************
* Fenced Code Blocks (gfm) *
******************************************************************/
// Find and convert gfm-inspired fenced code blocks into html.
Markdown.Extra.prototype.fencedCodeBlocks = function(text) {
function encodeCode(code) {
code = code.replace(/&/g, "&amp;");
code = code.replace(/</g, "&lt;");
code = code.replace(/>/g, "&gt;");
2013-06-03 22:19:52 +00:00
// These were escaped by PageDown before postNormalization
code = code.replace(/~D/g, "$$");
code = code.replace(/~T/g, "~");
2013-04-15 16:19:47 +00:00
return code;
}
var self = this;
text = text.replace(/(?:^|\n)```(.*)\n([\s\S]*?)\n```/g, function(match, m1, m2) {
var language = m1, codeblock = m2;
// adhere to specified options
var preclass = self.googleCodePrettify ? ' class="prettyprint"' : '';
var codeclass = '';
if (language) {
if (self.googleCodePrettify || self.highlightJs) {
// use html5 language- class names. supported by both prettify and highlight.js
codeclass = ' class="language-' + language + '"';
} else {
codeclass = ' class="' + language + '"';
}
}
var html = ['<pre', preclass, '><code', codeclass, '>',
encodeCode(codeblock), '</code></pre>'].join('');
// replace codeblock with placeholder until postConversion step
return self.hashExtraBlock(html);
});
return text;
};
/******************************************************************
* Definition Lists *
******************************************************************/
// Find and convert markdown extra definition lists into html.
Markdown.Extra.prototype.definitionLists = function(text) {
var wholeList = new RegExp(
['(\\x02\\n?|\\n\\n)' ,
'(?:' ,
'(' , // $1 = whole list
'(' , // $2
'[ ]{0,3}' ,
'((?:[ \\t]*\\S.*\\n)+)', // $3 = defined term
'\\n?' ,
'[ ]{0,3}:[ ]+' , // colon starting definition
')' ,
'([\\s\\S]+?)' ,
'(' , // $4
'(?=\\0x03)' , // \z
'|' ,
'(?=' ,
'\\n{2,}' ,
'(?=\\S)' ,
'(?!' , // Negative lookahead for another term
'[ ]{0,3}' ,
'(?:\\S.*\\n)+?' , // defined term
'\\n?' ,
'[ ]{0,3}:[ ]+' , // colon starting definition
')' ,
'(?!' , // Negative lookahead for another definition
'[ ]{0,3}:[ ]+' , // colon starting definition
')' ,
')' ,
')' ,
')' ,
')'
].join(''),
'gm'
);
var self = this;
text = addAnchors(text);
text = text.replace(wholeList, function(match, pre, list) {
var result = trim(self.processDefListItems(list));
result = "<dl>\n" + result + "\n</dl>";
return pre + self.hashExtraBlock(result) + "\n\n";
});
return removeAnchors(text);
};
// Process the contents of a single definition list, splitting it
// into individual term and definition list items.
Markdown.Extra.prototype.processDefListItems = function(listStr) {
var self = this;
var dt = new RegExp(
['(\\x02\\n?|\\n\\n+)' , // leading line
'(' , // definition terms = $1
'[ ]{0,3}' , // leading whitespace
'(?![:][ ]|[ ])' , // negative lookahead for a definition
// mark (colon) or more whitespace
'(?:\\S.*\\n)+?' , // actual term (not whitespace)
')' ,
'(?=\\n?[ ]{0,3}:[ ])' // lookahead for following line feed
].join(''), // with a definition mark
'gm'
);
var dd = new RegExp(
['\\n(\\n+)?' , // leading line = $1
'(' , // marker space = $2
'[ ]{0,3}' , // whitespace before colon
'[:][ ]+' , // definition mark (colon)
')' ,
'([\\s\\S]+?)' , // definition text = $3
'(?=\\n*' , // stop at next definition mark,
'(?:' , // next term or end of text
'\\n[ ]{0,3}[:][ ]|' ,
'<dt>|\\x03' , // \z
')' ,
')'
].join(''),
'gm'
);
listStr = addAnchors(listStr);
// trim trailing blank lines:
listStr = listStr.replace(/\n{2,}(?=\\x03)/, "\n");
// Process definition terms.
listStr = listStr.replace(dt, function(match, pre, termsStr) {
var terms = trim(termsStr).split("\n");
var text = '';
for (var i = 0; i < terms.length; i++) {
var term = terms[i];
// process spans inside dt
2013-06-02 19:35:44 +00:00
term = convertSpans(trim(term), self);
2013-04-15 16:19:47 +00:00
text += "\n<dt>" + term + "</dt>";
}
return text + "\n";
});
// Process actual definitions.
listStr = listStr.replace(dd, function(match, leadingLine, markerSpace, def) {
if (leadingLine || def.match(/\n{2,}/)) {
// replace marker with the appropriate whitespace indentation
def = Array(markerSpace.length + 1).join(' ') + def;
// process markdown inside definition
// TODO?: currently doesn't apply extensions
def = outdent(def) + "\n\n";
2013-06-02 19:35:44 +00:00
def = "\n" + convertAll(def, self) + "\n";
2013-04-15 16:19:47 +00:00
} else {
// convert span-level markdown inside definition
def = rtrim(def);
2013-06-02 19:35:44 +00:00
def = convertSpans(outdent(def), self);
2013-04-15 16:19:47 +00:00
}
return "\n<dd>" + def + "</dd>\n";
});
return removeAnchors(listStr);
};
})();