highlight.js

a RIP distance vector implementation for a simple graph

/* 
Syntax highlighting with language autodetection.  
http://softwaremaniacs.org/soft/highlight/
*/

var hljs = new function() {

  var DEFAULT_LANGUAGES = ['python', 'ruby', 'perl', 'php', 'css', 'xml', 'html', 'django', 'javascript', 'java', 'cpp', 'cs', 'sql', 'ini', 'diff'];
  var ALL_LANGUAGES = (DEFAULT_LANGUAGES.join(',') + ',' + ['1c', 'axapta', 'delphi', 'rib', 'rsl', 'vbscript', 'profile', 'dos', 'bash', 'lisp', 'smalltalk', 'mel'].join(',')).split(',');
  var LANGUAGE_GROUPS = {
    'xml': 'www',
    'html': 'www',
    'css': 'www',
    'django': 'www',
    'python': 'dynamic',
    'perl': 'dynamic',
    'php': 'dynamic',
    'ruby': 'dynamic',
    'cpp': 'static',
    'java': 'static',
    'delphi': 'static',
    'cs': 'static',
    'rib': 'renderman',
    'rsl': 'renderman'
  }

  var LANGUAGES = {}
  var selected_languages = {};
  
  function escape(value) {
    return value.replace(/&/gm, '&amp;').replace(/</gm, '&lt;').replace(/>/gm, '&gt;');
  }
  
  function contains(array, item) {
    if (!array)
      return false;
    for (var i = 0; i < array.length; i++)
      if (array[i] == item)
        return true;
    return false;
  }
  
  function highlight(language_name, value) {
    function compileSubModes(mode, language) {
      mode.sub_modes = [];
      for (var i = 0; i < mode.contains.length; i++) {
        for (var j = 0; j < language.modes.length; j++) {
          if (language.modes[j].className == mode.contains[i]) {
            mode.sub_modes[mode.sub_modes.length] = language.modes[j];
          }
        }
      }
    }
  
    function subMode(lexem, mode) {
      if (!mode.contains) {
        return null;
      }
      if (!mode.sub_modes) {
        compileSubModes(mode, language);
      }
      for (var i = 0; i < mode.sub_modes.length; i++) {
        if (mode.sub_modes[i].beginRe.test(lexem)) {
          return mode.sub_modes[i];
        }
      }
      return null;
    }
    
    function endOfMode(mode_index, lexem) {
      if (modes[mode_index].end && modes[mode_index].endRe.test(lexem))
        return 1;
      if (modes[mode_index].endsWithParent) {
        var level = endOfMode(mode_index - 1, lexem);
        return level ? level + 1 : 0;
      }
      return 0;
    }
    
    function isIllegal(lexem, mode) {
      return mode.illegalRe && mode.illegalRe.test(lexem);
    }
    
    function compileTerminators(mode, language) {
      var terminators = [];
      
      function addTerminator(re) {
        if (!contains(terminators, re)) {
          terminators[terminators.length] = re;
        }
      }
      
      if (mode.contains)
        for (var i = 0; i < language.modes.length; i++) {
          if (contains(mode.contains, language.modes[i].className)) {
            addTerminator(language.modes[i].begin);
          }
        }
      
      var index = modes.length - 1;
      do {
        if (modes[index].end) {
          addTerminator(modes[index].end);
        }
        index--;
      } while (modes[index + 1].endsWithParent);
      
      if (mode.illegal) {
        addTerminator(mode.illegal);
      }
      
      var terminator_re = '(' + terminators[0];
      for (var i = 0; i < terminators.length; i++)
        terminator_re += '|' + terminators[i];
      terminator_re += ')';
      return langRe(language, terminator_re);
    }

    function eatModeChunk(value, index) {
      var mode = modes[modes.length - 1];
      if (!mode.terminators) {
        mode.terminators = compileTerminators(mode, language);
      }
      value = value.substr(index);
      var match = mode.terminators.exec(value);
      if (!match) 
        return [value, '', true];
      if (match.index == 0)
        return ['', match[0], false];
      else
        return [value.substr(0, match.index), match[0], false];
    }
    
    function keywordMatch(mode, match) {
      var match_str = language.case_insensitive ? match[0].toLowerCase() : match[0]
      for (var className in mode.keywordGroups) {
        if (!mode.keywordGroups.hasOwnProperty(className))
          continue;
        var value = mode.keywordGroups[className].hasOwnProperty(match_str);
        if (value)
          return [className, value];
      }
      return false;
    }
    
    function processKeywords(buffer, mode) {
      if (!mode.keywords || !mode.lexems)
        return escape(buffer);
      if (!mode.lexemsRe) {
        var lexems_re = '(' + mode.lexems[0];
        for (var i = 1; i < mode.lexems.length; i++)
          lexems_re += '|' + mode.lexems[i];
        lexems_re += ')';
        mode.lexemsRe = langRe(language, lexems_re, true);
      }
      var result = '';
      var last_index = 0;
      mode.lexemsRe.lastIndex = 0;
      var match = mode.lexemsRe.exec(buffer);
      while (match) {
        result += escape(buffer.substr(last_index, match.index - last_index));
        keyword_match = keywordMatch(mode, match);
        if (keyword_match) {
          keyword_count += keyword_match[1];
          result += '<span class="'+ keyword_match[0] +'">' + escape(match[0]) + '</span>';
        } else {
          result += escape(match[0]);
        }
        last_index = mode.lexemsRe.lastIndex;
        match = mode.lexemsRe.exec(buffer);
      }
      result += escape(buffer.substr(last_index, buffer.length - last_index));
      return result;
    }
    
    function processBuffer(buffer, mode) {
      if (mode.subLanguage && selected_languages[mode.subLanguage]) {
        var result = highlight(mode.subLanguage, buffer);
        keyword_count += result.keyword_count;
        relevance += result.relevance;
        return result.value;
      } else {
        return processKeywords(buffer, mode);
      }
    }
    
    function startNewMode(mode, lexem) {
      if (mode.returnBegin) {
        result += '<span class="' + mode.className + '">';
        mode.buffer = '';
      } else if (mode.excludeBegin) {
        result += escape(lexem) + '<span class="' + mode.className + '">';
        mode.buffer = '';
      } else {
        result += '<span class="' + mode.className + '">';
        mode.buffer = lexem;
      }
      modes[modes.length] = mode;
    }
    
    function processModeInfo(buffer, lexem, end) {
      var current_mode = modes[modes.length - 1];
      if (end) {
        result += processBuffer(current_mode.buffer + buffer, current_mode);
        return false;
      }
      
      var new_mode = subMode(lexem, current_mode);
      if (new_mode) {
        result += processBuffer(current_mode.buffer + buffer, current_mode);
        startNewMode(new_mode, lexem);
        relevance += new_mode.relevance;
        return new_mode.returnBegin;
      }
      
      var end_level = endOfMode(modes.length - 1, lexem);
      if (end_level) {
        if (current_mode.returnEnd) {
          result += processBuffer(current_mode.buffer + buffer, current_mode) + '</span>';
        } else if (current_mode.excludeEnd) {
          result += processBuffer(current_mode.buffer + buffer, current_mode) + '</span>' + escape(lexem);
        } else {
          result += processBuffer(current_mode.buffer + buffer + lexem, current_mode) + '</span>';
        }
        while (end_level > 1) {
          result += '</span>';
          end_level--;
          modes.length--;
        }
        modes.length--;
        modes[modes.length - 1].buffer = '';
        if (current_mode.starts) {
          for (var i = 0; i < language.modes.length; i++) {
            if (language.modes[i].className == current_mode.starts) {
              startNewMode(language.modes[i], '');
              break;
            }
          }
        }
        return current_mode.returnEnd;
      }
      
      if (isIllegal(lexem, current_mode))
        throw 'Illegal';
    }
    
    var language = LANGUAGES[language_name];
    var modes = [language.defaultMode];
    var relevance = 0;
    var keyword_count = 0;
    var result = '';
    try {
      var index = 0;
      language.defaultMode.buffer = '';
      do {
        var mode_info = eatModeChunk(value, index);
        var return_lexem = processModeInfo(mode_info[0], mode_info[1], mode_info[2]);
        index += mode_info[0].length;
        if (!return_lexem) {