Subversion Repositories Integrator Subversion

// Released under MIT license

// Copyright (c) 2009-2010 Dominic Baggott

// Copyright (c) 2009-2010 Ash Berlin

// Copyright (c) 2011 Christoph Dorn <christoph@christophdorn.com> (http://www.christophdorn.com)

(function( expose ) {

/**

 *  class Markdown

 *

 *  Markdown processing in Javascript done right. We have very particular views

 *  on what constitutes 'right' which include:

 *

 *  - produces well-formed HTML (this means that em and strong nesting is

 *    important)

 *

 *  - has an intermediate representation to allow processing of parsed data (We

 *    in fact have two, both as [JsonML]: a markdown tree and an HTML tree).

 *

 *  - is easily extensible to add new dialects without having to rewrite the

 *    entire parsing mechanics

 *

 *  - has a good test suite

 *

 *  This implementation fulfills all of these (except that the test suite could

 *  do with expanding to automatically run all the fixtures from other Markdown

 *  implementations.)

 *

 *  ##### Intermediate Representation

 *

 *  *TODO* Talk about this :) Its JsonML, but document the node names we use.

 *

 *  [JsonML]: http://jsonml.org/ "JSON Markup Language"

 **/

var Markdown = expose.Markdown = function Markdown(dialect) {

  switch (typeof dialect) {

    case "undefined":

      this.dialect = Markdown.dialects.Gruber;

      break;

    case "object":

      this.dialect = dialect;

      break;

    default:

      if (dialect in Markdown.dialects) {

        this.dialect = Markdown.dialects[dialect];

      }

      else {

        throw new Error("Unknown Markdown dialect '" + String(dialect) + "'");

      }

      break;

  }

  this.em_state = [];

  this.strong_state = [];

  this.debug_indent = "";

};

/**

 *  parse( markdown, [dialect] ) -> JsonML

 *  - markdown (String): markdown string to parse

 *  - dialect (String | Dialect): the dialect to use, defaults to gruber

 *

 *  Parse `markdown` and return a markdown document as a Markdown.JsonML tree.

 **/

expose.parse = function( source, dialect ) {

  // dialect will default if undefined

  var md = new Markdown( dialect );

  return md.toTree( source );

};

/**

 *  toHTML( markdown, [dialect]  ) -> String

 *  toHTML( md_tree ) -> String

 *  - markdown (String): markdown string to parse

 *  - md_tree (Markdown.JsonML): parsed markdown tree

 *

 *  Take markdown (either as a string or as a JsonML tree) and run it through

 *  [[toHTMLTree]] then turn it into a well-formated HTML fragment.

 **/

expose.toHTML = function toHTML( source , dialect , options ) {

  var input = expose.toHTMLTree( source , dialect , options );

  return expose.renderJsonML( input );

};

/**

 *  toHTMLTree( markdown, [dialect] ) -> JsonML

 *  toHTMLTree( md_tree ) -> JsonML

 *  - markdown (String): markdown string to parse

 *  - dialect (String | Dialect): the dialect to use, defaults to gruber

 *  - md_tree (Markdown.JsonML): parsed markdown tree

 *

 *  Turn markdown into HTML, represented as a JsonML tree. If a string is given

 *  to this function, it is first parsed into a markdown tree by calling

 *  [[parse]].

 **/

expose.toHTMLTree = function toHTMLTree( input, dialect , options ) {

  // convert string input to an MD tree

  if ( typeof input ==="string" ) input = this.parse( input, dialect );

  // Now convert the MD tree to an HTML tree

  // remove references from the tree

  var attrs = extract_attr( input ),

      refs = {};

  if ( attrs && attrs.references ) {

    refs = attrs.references;

  }

  var html = convert_tree_to_html( input, refs , options );

  merge_text_nodes( html );

  return html;

};

// For Spidermonkey based engines

function mk_block_toSource() {

  return "Markdown.mk_block( " +

          uneval(this.toString()) +

          ", " +

          uneval(this.trailing) +

          ", " +

          uneval(this.lineNumber) +

          " )";

}

// node

function mk_block_inspect() {

  var util = require('util');

  return "Markdown.mk_block( " +

          util.inspect(this.toString()) +

          ", " +

          util.inspect(this.trailing) +

          ", " +

          util.inspect(this.lineNumber) +

          " )";

}

var mk_block = Markdown.mk_block = function(block, trail, line) {

  // Be helpful for default case in tests.

  if ( arguments.length == 1 ) trail = "\n\n";

  var s = new String(block);

  s.trailing = trail;

  // To make it clear its not just a string

  s.inspect = mk_block_inspect;

  s.toSource = mk_block_toSource;

  if (line != undefined)

    s.lineNumber = line;

  return s;

};

function count_lines( str ) {

  var n = 0, i = -1;

  while ( ( i = str.indexOf('\n', i+1) ) !== -1) n++;

  return n;

}

// Internal - split source into rough blocks

Markdown.prototype.split_blocks = function splitBlocks( input, startLine ) {

  // [\s\S] matches _anything_ (newline or space)

  var re = /([\s\S]+?)($|\n(?:\s*\n|$)+)/g,

      blocks = [],

      m;

  var line_no = 1;

  if ( ( m = /^(\s*\n)/.exec(input) ) != null ) {

    // skip (but count) leading blank lines

    line_no += count_lines( m[0] );

    re.lastIndex = m[0].length;

  }

  while ( ( m = re.exec(input) ) !== null ) {

    blocks.push( mk_block( m[1], m[2], line_no ) );

    line_no += count_lines( m[0] );

  }

  return blocks;

};

/**

 *  Markdown#processBlock( block, next ) -> undefined | [ JsonML, ... ]

 *  - block (String): the block to process

 *  - next (Array): the following blocks

 *

 * Process `block` and return an array of JsonML nodes representing `block`.

 *

 * It does this by asking each block level function in the dialect to process

 * the block until one can. Succesful handling is indicated by returning an

 * array (with zero or more JsonML nodes), failure by a false value.

 *

 * Blocks handlers are responsible for calling [[Markdown#processInline]]

 * themselves as appropriate.

 *

 * If the blocks were split incorrectly or adjacent blocks need collapsing you

 * can adjust `next` in place using shift/splice etc.

 *

 * If any of this default behaviour is not right for the dialect, you can

 * define a `__call__` method on the dialect that will get invoked to handle

 * the block processing.

 */

Markdown.prototype.processBlock = function processBlock( block, next ) {

  var cbs = this.dialect.block,

      ord = cbs.__order__;

  if ( "__call__" in cbs ) {

    return cbs.__call__.call(this, block, next);

  }

  for ( var i = 0; i < ord.length; i++ ) {

    //D:this.debug( "Testing", ord[i] );

    var res = cbs[ ord[i] ].call( this, block, next );

    if ( res ) {

      //D:this.debug("  matched");

      if ( !isArray(res) || ( res.length > 0 && !( isArray(res[0]) ) ) )

        this.debug(ord[i], "didn't return a proper array");

      //D:this.debug( "" );

      return res;

    }

  }

  // Uhoh! no match! Should we throw an error?

  return [];

};

Markdown.prototype.processInline = function processInline( block ) {

  return this.dialect.inline.__call__.call( this, String( block ) );

};

/**

 *  Markdown#toTree( source ) -> JsonML

 *  - source (String): markdown source to parse

 *

 *  Parse `source` into a JsonML tree representing the markdown document.

 **/

// custom_tree means set this.tree to `custom_tree` and restore old value on return

Markdown.prototype.toTree = function toTree( source, custom_root ) {

  var blocks = source instanceof Array ? source : this.split_blocks( source );

  // Make tree a member variable so its easier to mess with in extensions

  var old_tree = this.tree;

  try {

    this.tree = custom_root || this.tree || [ "markdown" ];

    blocks:

    while ( blocks.length ) {

      var b = this.processBlock( blocks.shift(), blocks );

      // Reference blocks and the like won't return any content

      if ( !b.length ) continue blocks;

      this.tree.push.apply( this.tree, b );

    }

    return this.tree;

  }

  finally {

    if ( custom_root ) {

      this.tree = old_tree;

    }

  }

};

// Noop by default

Markdown.prototype.debug = function () {

  var args = Array.prototype.slice.call( arguments);

  args.unshift(this.debug_indent);

  if (typeof print !== "undefined")

      print.apply( print, args );

  if (typeof console !== "undefined" && typeof console.log !== "undefined")

      console.log.apply( null, args );

}

Markdown.prototype.loop_re_over_block = function( re, block, cb ) {

  // Dont use /g regexps with this

  var m,

      b = block.valueOf();

  while ( b.length && (m = re.exec(b) ) != null) {

    b = b.substr( m[0].length );

    cb.call(this, m);

  }

  return b;

};

/**

 * Markdown.dialects

 *

 * Namespace of built-in dialects.

 **/

Markdown.dialects = {};

/**

 * Markdown.dialects.Gruber

 *

 * The default dialect that follows the rules set out by John Gruber's

 * markdown.pl as closely as possible. Well actually we follow the behaviour of

 * that script which in some places is not exactly what the syntax web page

 * says.

 **/

Markdown.dialects.Gruber = {

  block: {

    atxHeader: function atxHeader( block, next ) {

      var m = block.match( /^(#{1,6})\s*(.*?)\s*#*\s*(?:\n|$)/ );

      if ( !m ) return undefined;

      var header = [ "header", { level: m[ 1 ].length } ];

      Array.prototype.push.apply(header, this.processInline(m[ 2 ]));

      if ( m[0].length < block.length )

        next.unshift( mk_block( block.substr( m[0].length ), block.trailing, block.lineNumber + 2 ) );

      return [ header ];

    },

    setextHeader: function setextHeader( block, next ) {

      var m = block.match( /^(.*)\n([-=])\2\2+(?:\n|$)/ );

      if ( !m ) return undefined;

      var level = ( m[ 2 ] === "=" ) ? 1 : 2;

      var header = [ "header", { level : level }, m[ 1 ] ];

      if ( m[0].length < block.length )

        next.unshift( mk_block( block.substr( m[0].length ), block.trailing, block.lineNumber + 2 ) );

      return [ header ];

    },

    code: function code( block, next ) {

      // |    Foo

      // |bar

      // should be a code block followed by a paragraph. Fun

      //

      // There might also be adjacent code block to merge.

      var ret = [],

          re = /^(?: {0,3}\t| {4})(.*)\n?/,

          lines;

      // 4 spaces + content

      if ( !block.match( re ) ) return undefined;

      block_search:

      do {

        // Now pull out the rest of the lines

        var b = this.loop_re_over_block(

                  re, block.valueOf(), function( m ) { ret.push( m[1] ); } );

        if (b.length) {

          // Case alluded to in first comment. push it back on as a new block

          next.unshift( mk_block(b, block.trailing) );

          break block_search;

        }

        else if (next.length) {

          // Check the next block - it might be code too

          if ( !next[0].match( re ) ) break block_search;

          // Pull how how many blanks lines follow - minus two to account for .join

          ret.push ( block.trailing.replace(/[^\n]/g, '').substring(2) );

          block = next.shift();

        }

        else {

          break block_search;

        }

      } while (true);

      return [ [ "code_block", ret.join("\n") ] ];

    },

    horizRule: function horizRule( block, next ) {

      // this needs to find any hr in the block to handle abutting blocks

      var m = block.match( /^(?:([\s\S]*?)\n)?[ \t]*([-_*])(?:[ \t]*\2){2,}[ \t]*(?:\n([\s\S]*))?$/ );

      if ( !m ) {

        return undefined;

      }

      var jsonml = [ [ "hr" ] ];

      // if there's a leading abutting block, process it

      if ( m[ 1 ] ) {

        jsonml.unshift.apply( jsonml, this.processBlock( m[ 1 ], [] ) );

      }

      // if there's a trailing abutting block, stick it into next

      if ( m[ 3 ] ) {

        next.unshift( mk_block( m[ 3 ] ) );

      }

      return jsonml;

    },

    // There are two types of lists. Tight and loose. Tight lists have no whitespace

    // between the items (and result in text just in the <li>) and loose lists,

    // which have an empty line between list items, resulting in (one or more)

    // paragraphs inside the <li>.

    //

    // There are all sorts weird edge cases about the original markdown.pl's

    // handling of lists:

    //

    // * Nested lists are supposed to be indented by four chars per level. But

    //   if they aren't, you can get a nested list by indenting by less than

    //   four so long as the indent doesn't match an indent of an existing list

    //   item in the 'nest stack'.

    //

    // * The type of the list (bullet or number) is controlled just by the

    //    first item at the indent. Subsequent changes are ignored unless they

    //    are for nested lists

    //

    lists: (function( ) {

      // Use a closure to hide a few variables.

      var any_list = "[*+-]|\\d+\\.",

          bullet_list = /[*+-]/,

          number_list = /\d+\./,

          // Capture leading indent as it matters for determining nested lists.

          is_list_re = new RegExp( "^( {0,3})(" + any_list + ")[ \t]+" ),

          indent_re = "(?: {0,3}\\t| {4})";

      // TODO: Cache this regexp for certain depths.

      // Create a regexp suitable for matching an li for a given stack depth

      function regex_for_depth( depth ) {

        return new RegExp(

          // m[1] = indent, m[2] = list_type

          "(?:^(" + indent_re + "{0," + depth + "} {0,3})(" + any_list + ")\\s+)|" +

          // m[3] = cont

          "(^" + indent_re + "{0," + (depth-1) + "}[ ]{0,4})"

        );

      }

      function expand_tab( input ) {

        return input.replace( / {0,3}\t/g, "    " );

      }

      // Add inline content `inline` to `li`. inline comes from processInline

      // so is an array of content

      function add(li, loose, inline, nl) {

        if (loose) {

          li.push( [ "para" ].concat(inline) );

          return;

        }

        // Hmmm, should this be any block level element or just paras?

        var add_to = li[li.length -1] instanceof Array && li[li.length - 1][0] == "para"

                   ? li[li.length -1]

                   : li;

        // If there is already some content in this list, add the new line in

        if (nl && li.length > 1) inline.unshift(nl);

        for (var i=0; i < inline.length; i++) {

          var what = inline[i],

              is_str = typeof what == "string";

          if (is_str && add_to.length > 1 && typeof add_to[add_to.length-1] == "string" ) {

            add_to[ add_to.length-1 ] += what;

          }

          else {

            add_to.push( what );

          }

        }

      }

      // contained means have an indent greater than the current one. On

      // *every* line in the block

      function get_contained_blocks( depth, blocks ) {

        var re = new RegExp( "^(" + indent_re + "{" + depth + "}.*?\\n?)*$" ),

            replace = new RegExp("^" + indent_re + "{" + depth + "}", "gm"),

            ret = [];

        while ( blocks.length > 0 ) {

          if ( re.exec( blocks[0] ) ) {

            var b = blocks.shift(),

                // Now remove that indent

                x = b.replace( replace, "");

            ret.push( mk_block( x, b.trailing, b.lineNumber ) );

          }

          break;

        }

        return ret;

      }

      // passed to stack.forEach to turn list items up the stack into paras

      function paragraphify(s, i, stack) {

        var list = s.list;

        var last_li = list[list.length-1];

        if (last_li[1] instanceof Array && last_li[1][0] == "para") {

          return;

        }

        if (i+1 == stack.length) {

          // Last stack frame

          // Keep the same array, but replace the contents

          last_li.push( ["para"].concat( last_li.splice(1) ) );

        }

        else {

          var sublist = last_li.pop();

          last_li.push( ["para"].concat( last_li.splice(1) ), sublist );

        }

      }

      // The matcher function

      return function( block, next ) {

        var m = block.match( is_list_re );

        if ( !m ) return undefined;

        function make_list( m ) {

          var list = bullet_list.exec( m[2] )

                   ? ["bulletlist"]

                   : ["numberlist"];

          stack.push( { list: list, indent: m[1] } );

          return list;

        }

        var stack = [], // Stack of lists for nesting.

            list = make_list( m ),

            last_li,

            loose = false,

            ret = [ stack[0].list ],

            i;

        // Loop to search over block looking for inner block elements and loose lists

        loose_search:

        while( true ) {

          // Split into lines preserving new lines at end of line

          var lines = block.split( /(?=\n)/ );

          // We have to grab all lines for a li and call processInline on them

          // once as there are some inline things that can span lines.

          var li_accumulate = "";

          // Loop over the lines in this block looking for tight lists.

          tight_search:

          for (var line_no=0; line_no < lines.length; line_no++) {

            var nl = "",

                l = lines[line_no].replace(/^\n/, function(n) { nl = n; return ""; });

            // TODO: really should cache this

            var line_re = regex_for_depth( stack.length );

            m = l.match( line_re );

            //print( "line:", uneval(l), "\nline match:", uneval(m) );

            // We have a list item

            if ( m[1] !== undefined ) {

              // Process the previous list item, if any

              if ( li_accumulate.length ) {

                add( last_li, loose, this.processInline( li_accumulate ), nl );

                // Loose mode will have been dealt with. Reset it

                loose = false;

                li_accumulate = "";

              }

              m[1] = expand_tab( m[1] );

              var wanted_depth = Math.floor(m[1].length/4)+1;

              //print( "want:", wanted_depth, "stack:", stack.length);

              if ( wanted_depth > stack.length ) {

                // Deep enough for a nested list outright

                //print ( "new nested list" );

                list = make_list( m );

                last_li.push( list );

                last_li = list[1] = [ "listitem" ];

              }

              else {

                // We aren't deep enough to be strictly a new level. This is

                // where Md.pl goes nuts. If the indent matches a level in the

                // stack, put it there, else put it one deeper then the

                // wanted_depth deserves.

                var found = false;

                for (i = 0; i < stack.length; i++) {

                  if ( stack[ i ].indent != m[1] ) continue;

                  list = stack[ i ].list;

                  stack.splice( i+1 );

                  found = true;

                  break;

                }

                if (!found) {

                  //print("not found. l:", uneval(l));

                  wanted_depth++;

                  if (wanted_depth <= stack.length) {

                    stack.splice(wanted_depth);

                    //print("Desired depth now", wanted_depth, "stack:", stack.length);

                    list = stack[wanted_depth-1].list;

                    //print("list:", uneval(list) );

                  }

                  else {

                    //print ("made new stack for messy indent");

                    list = make_list(m);

                    last_li.push(list);

                  }

                }

                //print( uneval(list), "last", list === stack[stack.length-1].list );

                last_li = [ "listitem" ];

                list.push(last_li);

              } // end depth of shenegains

              nl = "";

            }

            // Add content

            if (l.length > m[0].length) {

              li_accumulate += nl + l.substr( m[0].length );

            }

          } // tight_search

          if ( li_accumulate.length ) {

            add( last_li, loose, this.processInline( li_accumulate ), nl );

            // Loose mode will have been dealt with. Reset it

            loose = false;

            li_accumulate = "";

          }

          // Look at the next block - we might have a loose list. Or an extra

          // paragraph for the current li

          var contained = get_contained_blocks( stack.length, next );

          // Deal with code blocks or properly nested lists

          if (contained.length > 0) {

            // Make sure all listitems up the stack are paragraphs

            forEach( stack, paragraphify, this);

            last_li.push.apply( last_li, this.toTree( contained, [] ) );

          }

          var next_block = next[0] && next[0].valueOf() || "";

          if ( next_block.match(is_list_re) || next_block.match( /^ / ) ) {

            block = next.shift();

            // Check for an HR following a list: features/lists/hr_abutting

            var hr = this.dialect.block.horizRule( block, next );

            if (hr) {

              ret.push.apply(ret, hr);

              break;

            }

            // Make sure all listitems up the stack are paragraphs

            forEach( stack, paragraphify, this);

            loose = true;

            continue loose_search;

          }

          break;

        } // loose_search

        return ret;

      };

    })(),

    blockquote: function blockquote( block, next ) {

      if ( !block.match( /^>/m ) )

        return undefined;

      var jsonml = [];

      // separate out the leading abutting block, if any

      if ( block[ 0 ] != ">" ) {

        var lines = block.split( /\n/ ),

            prev = [];

        // keep shifting lines until you find a crotchet

        while ( lines.length && lines[ 0 ][ 0 ] != ">" ) {

            prev.push( lines.shift() );

        }

        // reassemble!

        block = lines.join( "\n" );

        jsonml.push.apply( jsonml, this.processBlock( prev.join( "\n" ), [] ) );

      }

      // if the next block is also a blockquote merge it in

      while ( next.length && next[ 0 ][ 0 ] == ">" ) {

        var b = next.shift();

        block = new String(block + block.trailing + b);

        block.trailing = b.trailing;

      }

      // Strip off the leading "> " and re-process as a block.

      var input = block.replace( /^> ?/gm, '' ),

          old_tree = this.tree;

      jsonml.push( this.toTree( input, [ "blockquote" ] ) );

      return jsonml;

    },

    referenceDefn: function referenceDefn( block, next) {

      var re = /^\s*\[(.*?)\]:\s*(\S+)(?:\s+(?:(['"])(.*?)\3|\((.*?)\)))?\n?/;

      // interesting matches are [ , ref_id, url, , title, title ]

      if ( !block.match(re) )

        return undefined;

      // make an attribute node if it doesn't exist

      if ( !extract_attr( this.tree ) ) {

        this.tree.splice( 1, 0, {} );

      }

      var attrs = extract_attr( this.tree );

      // make a references hash if it doesn't exist

      if ( attrs.references === undefined ) {

        attrs.references = {};

      }

      var b = this.loop_re_over_block(re, block, function( m ) {

        if ( m[2] && m[2][0] == '<' && m[2][m[2].length-1] == '>' )

          m[2] = m[2].substring( 1, m[2].length - 1 );

        var ref = attrs.references[ m[1].toLowerCase() ] = {

          href: m[2]

        };

        if (m[4] !== undefined)

          ref.title = m[4];

        else if (m[5] !== undefined)

          ref.title = m[5];

      } );

      if (b.length)

        next.unshift( mk_block( b, block.trailing ) );

      return [];

    },

    para: function para( block, next ) {

      // everything's a para!

      return [ ["para"].concat( this.processInline( block ) ) ];

    }

  }

};

Markdown.dialects.Gruber.inline = {

    __oneElement__: function oneElement( text, patterns_or_re, previous_nodes ) {

      var m,

          res,

          lastIndex = 0;

      patterns_or_re = patterns_or_re || this.dialect.inline.__patterns__;

      var re = new RegExp( "([\\s\\S]*?)(" + (patterns_or_re.source || patterns_or_re) + ")" );

      m = re.exec( text );

      if (!m) {

        // Just boring text

        return [ text.length, text ];

      }

      else if ( m[1] ) {

        // Some un-interesting text matched. Return that first