view release on metacpan or  search on metacpan

share/endpoint/www/js/tokenizejavascript.js  view on Meta::CPAN

/* Tokenizer for JavaScript code */

var tokenizeJavaScript = (function() {
  // Advance the stream until the given character (not preceded by a
  // backslash) is encountered, or the end of the line is reached.
  function nextUntilUnescaped(source, end) {
    var escaped = false;
    while (!source.endOfLine()) {
      var next = source.next();
      if (next == end && !escaped)
        return false;
      escaped = !escaped && next == "\\";
    }
    return escaped;
  }

  // A map of JavaScript's keywords. The a/b/c keyword distinction is
  // very rough, but it gives the parser enough information to parse
  // correct code correctly (we don't care that much how we parse
  // incorrect code). The style information included in these objects
  // is used by the highlighter to pick the correct CSS style for a
  // token.
  var keywords = function(){
    function result(type, style){
      return {type: type, style: "js-" + style};
    }
    // keywords that take a parenthised expression, and then a
    // statement (if)
    var keywordA = result("keyword a", "keyword");
    // keywords that take just a statement (else)
    var keywordB = result("keyword b", "keyword");
    // keywords that optionally take an expression, and form a
    // statement (return)
    var keywordC = result("keyword c", "keyword");
    var operator = result("operator", "keyword");
    var atom = result("atom", "atom");
    return {
      "if": keywordA, "while": keywordA, "with": keywordA,
      "else": keywordB, "do": keywordB, "try": keywordB, "finally": keywordB,
      "return": keywordC, "break": keywordC, "continue": keywordC, "new": keywordC, "delete": keywordC, "throw": keywordC,
      "in": operator, "typeof": operator, "instanceof": operator,
      "var": result("var", "keyword"), "function": result("function", "keyword"), "catch": result("catch", "keyword"),
      "for": result("for", "keyword"), "switch": result("switch", "keyword"),
      "case": result("case", "keyword"), "default": result("default", "keyword"),
      "true": atom, "false": atom, "null": atom, "undefined": atom, "NaN": atom, "Infinity": atom
    };
  }();

  // Some helper regexps
  var isOperatorChar = /[+\-*&%=<>!?|]/;
  var isHexDigit = /[0-9A-Fa-f]/;
  var isWordChar = /[\w\$_]/;

  // Wrapper around jsToken that helps maintain parser state (whether
  // we are inside of a multi-line comment and whether the next token
  // could be a regular expression).
  function jsTokenState(inside, regexp) {
    return function(source, setState) {
      var newInside = inside;
      var type = jsToken(inside, regexp, source, function(c) {newInside = c;});
      var newRegexp = type.type == "operator" || type.type == "keyword c" || type.type.match(/^[\[{}\(,;:]$/);
      if (newRegexp != regexp || newInside != inside)
        setState(jsTokenState(newInside, newRegexp));
      return type;
    };
  }

  // The token reader, intended to be used by the tokenizer from
  // tokenize.js (through jsTokenState). Advances the source stream
  // over a token, and returns an object containing the type and style
  // of that token.
  function jsToken(inside, regexp, source, setInside) {
    function readHexNumber(){
      source.next(); // skip the 'x'
      source.nextWhileMatches(isHexDigit);
      return {type: "number", style: "js-atom"};
    }

    function readNumber() {
      source.nextWhileMatches(/[0-9]/);
      if (source.equals(".")){
        source.next();
        source.nextWhileMatches(/[0-9]/);
      }
      if (source.equals("e") || source.equals("E")){
        source.next();
        if (source.equals("-"))
          source.next();
        source.nextWhileMatches(/[0-9]/);
      }
      return {type: "number", style: "js-atom"};
    }
    // Read a word, look it up in keywords. If not found, it is a
    // variable, otherwise it is a keyword of the type found.
    function readWord() {
      source.nextWhileMatches(isWordChar);
      var word = source.get();
      var known = keywords.hasOwnProperty(word) && keywords.propertyIsEnumerable(word) && keywords[word];
      return known ? {type: known.type, style: known.style, content: word} :
      {type: "variable", style: "js-variable", content: word};
    }
    function readRegexp() {
      nextUntilUnescaped(source, "/");
      source.nextWhileMatches(/[gi]/);
      return {type: "regexp", style: "js-string"};