style-script 1.0.0

data/lib/style_script/lexer.js

@@ -0,0 +1,363 @@
+(function(){
+  var ASSIGNMENT, CALLABLE, CODE, COMMENT, COMMENT_CLEANER, HEREDOC, HEREDOC_INDENT, IDENTIFIER, JS, JS_CLEANER, KEYWORDS, LAST_DENT, LAST_DENTS, MULTILINER, MULTI_DENT, NOT_REGEX, NO_NEWLINE, NUMBER, OPERATOR, REGEX, Rewriter, STRING, STRING_NEWLINES, WHITESPACE, lex;
+  Rewriter = require('./rewriter').Rewriter;
+  // The lexer reads a stream of StyleScript and divvys it up into tagged
+  // tokens. A minor bit of the ambiguity in the grammar has been avoided by
+  // pushing some extra smarts into the Lexer.
+  exports.Lexer = (lex = function lex() {  });
+  // Constants ============================================================
+  // The list of keywords passed verbatim to the parser.
+  KEYWORDS = ["if", "else", "then", "unless", "true", "false", "yes", "no", "on", "off", "and", "or", "is", "isnt", "not", "new", "return", "arguments", "try", "catch", "finally", "throw", "break", "continue", "for", "in", "of", "by", "where", "while", "delete", "instanceof", "typeof", "switch", "when", "super", "extends"];
+  // Token matching regexes.
+  IDENTIFIER = /^([a-zA-Z$_](\w|\$)*)/;
+  NUMBER = /^(\b((0(x|X)[0-9a-fA-F]+)|([0-9]+(\.[0-9]+)?(e[+\-]?[0-9]+)?)))\b/i;
+  STRING = /^(""|''|"([\s\S]*?)([^\\]|\\\\)"|'([\s\S]*?)([^\\]|\\\\)')/;
+  HEREDOC = /^("{6}|'{6}|"{3}\n?([\s\S]*?)\n?([ \t]*)"{3}|'{3}\n?([\s\S]*?)\n?([ \t]*)'{3})/;
+  JS = /^(``|`([\s\S]*?)([^\\]|\\\\)`)/;
+  OPERATOR = /^([+\*&|\/\-%=<>:!?]+)/;
+  WHITESPACE = /^([ \t]+)/;
+  COMMENT = /^(((\n?[ \t]*)?#[^\n]*)+)/;
+  CODE = /^((-|=)>)/;
+  REGEX = /^(\/(.*?)([^\\]|\\\\)\/[imgy]{0,4})/;
+  MULTI_DENT = /^((\n([ \t]*))+)(\.)?/;
+  LAST_DENTS = /\n([ \t]*)/g;
+  LAST_DENT = /\n([ \t]*)/;
+  ASSIGNMENT = /^(:|=)$/;
+  // Token cleaning regexes.
+  JS_CLEANER = /(^`|`$)/g;
+  MULTILINER = /\n/g;
+  STRING_NEWLINES = /\n[ \t]*/g;
+  COMMENT_CLEANER = /(^[ \t]*#|\n[ \t]*$)/mg;
+  NO_NEWLINE = /^([+\*&|\/\-%=<>:!.\\][<>=&|]*|and|or|is|isnt|not|delete|typeof|instanceof)$/;
+  HEREDOC_INDENT = /^[ \t]+/g;
+  // Tokens which a regular expression will never immediately follow, but which
+  // a division operator might.
+  // See: http://www.mozilla.org/js/language/js20-2002-04/rationale/syntax.html#regular-expressions
+  NOT_REGEX = ['IDENTIFIER', 'NUMBER', 'REGEX', 'STRING', ')', '++', '--', ']', '}', 'FALSE', 'NULL', 'TRUE'];
+  // Tokens which could legitimately be invoked or indexed.
+  CALLABLE = ['IDENTIFIER', 'SUPER', ')', ']', '}', 'STRING'];
+  // Scan by attempting to match tokens one character at a time. Slow and steady.
+  lex.prototype.tokenize = function tokenize(code) {
+    this.code = code;
+    // Cleanup code by remove extra line breaks, TODO: chomp
+    this.i = 0;
+    // Current character position we're parsing
+    this.line = 1;
+    // The current line.
+    this.indent = 0;
+    // The current indent level.
+    this.indents = [];
+    // The stack of all indent levels we are currently within.
+    this.tokens = [];
+    // Collection of all parsed tokens in the form [:TOKEN_TYPE, value]
+    this.spaced = null;
+    // The last token that has a space following it.
+    while (this.i < this.code.length) {
+      this.chunk = this.code.slice(this.i);
+      this.extract_next_token();
+    }
+    this.close_indentation();
+    return (new Rewriter()).rewrite(this.tokens);
+  };
+  // At every position, run through this list of attempted matches,
+  // short-circuiting if any of them succeed.
+  lex.prototype.extract_next_token = function extract_next_token() {
+    if (this.identifier_token()) {
+      return null;
+    }
+    if (this.number_token()) {
+      return null;
+    }
+    if (this.heredoc_token()) {
+      return null;
+    }
+    if (this.string_token()) {
+      return null;
+    }
+    if (this.js_token()) {
+      return null;
+    }
+    if (this.regex_token()) {
+      return null;
+    }
+    if (this.indent_token()) {
+      return null;
+    }
+    if (this.comment_token()) {
+      return null;
+    }
+    if (this.whitespace_token()) {
+      return null;
+    }
+    return this.literal_token();
+  };
+  // Tokenizers ==========================================================
+  // Matches identifying literals: variables, keywords, method names, etc.
+  lex.prototype.identifier_token = function identifier_token() {
+    var id, tag;
+    if (!((id = this.match(IDENTIFIER, 1)))) {
+      return false;
+    }
+    // Keywords are special identifiers tagged with their own name,
+    // 'if' will result in an ['IF', "if"] token.
+    tag = KEYWORDS.indexOf(id) >= 0 ? id.toUpperCase() : 'IDENTIFIER';
+    if (tag === 'WHEN' && (this.tag() === 'OUTDENT' || this.tag() === 'INDENT')) {
+      tag = 'LEADING_WHEN';
+    }
+    if (tag === 'IDENTIFIER' && this.value() === '::') {
+      this.tag(-1, 'PROTOTYPE_ACCESS');
+    }
+    if (tag === 'IDENTIFIER' && this.value() === '.' && !(this.value(2) === '.')) {
+      if (this.tag(2) === '?') {
+        this.tag(1, 'SOAK_ACCESS');
+        this.tokens.splice(-2, 1);
+      } else {
+        this.tag(1, 'PROPERTY_ACCESS');
+      }
+    }
+    this.token(tag, id);
+    this.i += id.length;
+    return true;
+  };
+  // Matches numbers, including decimals, hex, and exponential notation.
+  lex.prototype.number_token = function number_token() {
+    var number;
+    if (!((number = this.match(NUMBER, 1)))) {
+      return false;
+    }
+    this.token('NUMBER', number);
+    this.i += number.length;
+    return true;
+  };
+  // Matches strings, including multi-line strings.
+  lex.prototype.string_token = function string_token() {
+    var escaped, string;
+    if (!((string = this.match(STRING, 1)))) {
+      return false;
+    }
+    escaped = string.replace(STRING_NEWLINES, " \\\n");
+    this.token('STRING', escaped);
+    this.line += this.count(string, "\n");
+    this.i += string.length;
+    return true;
+  };
+  // Matches heredocs, adjusting indentation to the correct level.
+  lex.prototype.heredoc_token = function heredoc_token() {
+    var doc, indent, match;
+    if (!((match = this.chunk.match(HEREDOC)))) {
+      return false;
+    }
+    doc = match[2] || match[4];
+    indent = doc.match(HEREDOC_INDENT).sort()[0];
+    doc = doc.replace(new RegExp("^" + indent, 'g'), '').replace(MULTILINER, "\\n").replace('"', '\\"');
+    this.token('STRING', '"' + doc + '"');
+    this.line += this.count(match[1], "\n");
+    this.i += match[1].length;
+    return true;
+  };
+  // Matches interpolated JavaScript.
+  lex.prototype.js_token = function js_token() {
+    var script;
+    if (!((script = this.match(JS, 1)))) {
+      return false;
+    }
+    this.token('JS', script.replace(JS_CLEANER, ''));
+    this.i += script.length;
+    return true;
+  };
+  // Matches regular expression literals.
+  lex.prototype.regex_token = function regex_token() {
+    var regex;
+    if (!((regex = this.match(REGEX, 1)))) {
+      return false;
+    }
+    if (NOT_REGEX.indexOf(this.tag()) >= 0) {
+      return false;
+    }
+    this.token('REGEX', regex);
+    this.i += regex.length;
+    return true;
+  };
+  // Matches and conumes comments.
+  lex.prototype.comment_token = function comment_token() {
+    var comment;
+    if (!((comment = this.match(COMMENT, 1)))) {
+      return false;
+    }
+    this.line += (comment.match(MULTILINER) || []).length;
+    this.token('COMMENT', comment.replace(COMMENT_CLEANER, '').split(MULTILINER));
+    this.token('TERMINATOR', "\n");
+    this.i += comment.length;
+    return true;
+  };
+  // Record tokens for indentation differing from the previous line.
+  lex.prototype.indent_token = function indent_token() {
+    var diff, indent, next_character, no_newlines, size;
+    if (!((indent = this.match(MULTI_DENT, 1)))) {
+      return false;
+    }
+    this.line += indent.match(MULTILINER).length;
+    this.i += indent.length;
+    next_character = this.chunk.match(MULTI_DENT)[4];
+    no_newlines = next_character === '.' || (this.value().match(NO_NEWLINE) && this.tokens[this.tokens.length - 2][0] !== '.' && !this.value().match(CODE));
+    if (no_newlines) {
+      return this.suppress_newlines(indent);
+    }
+    size = indent.match(LAST_DENTS).reverse()[0].match(LAST_DENT)[1].length;
+    if (size === this.indent) {
+      return this.newline_token(indent);
+    }
+    if (size > this.indent) {
+      diff = size - this.indent;
+      this.token('INDENT', diff);
+      this.indents.push(diff);
+    } else {
+      this.outdent_token(this.indent - size);
+    }
+    this.indent = size;
+    return true;
+  };
+  // Record an oudent token or tokens, if we're moving back inwards past
+  // multiple recorded indents.
+  lex.prototype.outdent_token = function outdent_token(move_out) {
+    var last_indent;
+    while (move_out > 0 && this.indents.length) {
+      last_indent = this.indents.pop();
+      this.token('OUTDENT', last_indent);
+      move_out -= last_indent;
+    }
+    this.token('TERMINATOR', "\n");
+    return true;
+  };
+  // Matches and consumes non-meaningful whitespace.
+  lex.prototype.whitespace_token = function whitespace_token() {
+    var space;
+    if (!((space = this.match(WHITESPACE, 1)))) {
+      return false;
+    }
+    this.spaced = this.value();
+    this.i += space.length;
+    return true;
+  };
+  // Multiple newlines get merged together.
+  // Use a trailing \ to escape newlines.
+  lex.prototype.newline_token = function newline_token(newlines) {
+    if (!(this.value() === "\n")) {
+      this.token('TERMINATOR', "\n");
+    }
+    return true;
+  };
+  // Tokens to explicitly escape newlines are removed once their job is done.
+  lex.prototype.suppress_newlines = function suppress_newlines(newlines) {
+    if (this.value() === "\\") {
+      this.tokens.pop();
+    }
+    return true;
+  };
+  // We treat all other single characters as a token. Eg.: ( ) , . !
+  // Multi-character operators are also literal tokens, so that Racc can assign
+  // the proper order of operations.
+  lex.prototype.literal_token = function literal_token() {
+    var match, tag, value;
+    match = this.chunk.match(OPERATOR);
+    value = match && match[1];
+    if (value && value.match(CODE)) {
+      this.tag_parameters();
+    }
+    value = value || this.chunk.substr(0, 1);
+    tag = value.match(ASSIGNMENT) ? 'ASSIGN' : value;
+    if (value === ';') {
+      tag = 'TERMINATOR';
+    }
+    if (this.value() !== this.spaced && CALLABLE.indexOf(this.tag()) >= 0) {
+      if (value === '(') {
+        tag = 'CALL_START';
+      }
+      if (value === '[') {
+        tag = 'INDEX_START';
+      }
+    }
+    this.token(tag, value);
+    this.i += value.length;
+    return true;
+  };
+  // Helpers =============================================================
+  // Add a token to the results, taking note of the line number.
+  lex.prototype.token = function token(tag, value) {
+    return this.tokens.push([tag, value]);
+    // this.tokens.push([tag, Value.new(value, @line)])
+  };
+  // Look at a tag in the current token stream.
+  lex.prototype.tag = function tag(index, tag) {
+    var tok;
+    if (!((tok = this.tokens[this.tokens.length - (index || 1)]))) {
+      return null;
+    }
+    if ((typeof tag !== "undefined" && tag !== null)) {
+      return (tok[0] = tag);
+    }
+    return tok[0];
+  };
+  // Look at a value in the current token stream.
+  lex.prototype.value = function value(index, val) {
+    var tok;
+    if (!((tok = this.tokens[this.tokens.length - (index || 1)]))) {
+      return null;
+    }
+    if ((typeof val !== "undefined" && val !== null)) {
+      return (tok[1] = val);
+    }
+    return tok[1];
+  };
+  // Count the occurences of a character in a string.
+  lex.prototype.count = function count(string, letter) {
+    var num, pos;
+    num = 0;
+    pos = string.indexOf(letter);
+    while (pos !== -1) {
+      count += 1;
+      pos = string.indexOf(letter, pos + 1);
+    }
+    return count;
+  };
+  // Attempt to match a string against the current chunk, returning the indexed
+  // match.
+  lex.prototype.match = function match(regex, index) {
+    var m;
+    if (!((m = this.chunk.match(regex)))) {
+      return false;
+    }
+    return m ? m[index] : false;
+  };
+  // A source of ambiguity in our grammar was parameter lists in function
+  // definitions (as opposed to argument lists in function calls). Tag
+  // parameter identifiers in order to avoid this. Also, parameter lists can
+  // make use of splats.
+  lex.prototype.tag_parameters = function tag_parameters() {
+    var i, tok;
+    if (this.tag() !== ')') {
+      return null;
+    }
+    i = 0;
+    while (true) {
+      i += 1;
+      tok = this.tokens[this.tokens.length - i];
+      if (!tok) {
+        return null;
+      }
+      if (tok[0] === 'IDENTIFIER') {
+        tok[0] = 'PARAM';
+      } else if (tok[0] === ')') {
+        tok[0] = 'PARAM_END';
+      } else if (tok[0] === '(') {
+        return (tok[0] = 'PARAM_START');
+      }
+    }
+    return true;
+  };
+  // Close up all remaining open blocks. IF the first token is an indent,
+  // axe it.
+  lex.prototype.close_indentation = function close_indentation() {
+    return this.outdent_token(this.indent);
+  };
+})();

data/lib/style_script/lexer.rb

@@ -0,0 +1,272 @@
+module StyleScript
+
+  # The lexer reads a stream of StyleScript and divvys it up into tagged
+  # tokens. A minor bit of the ambiguity in the grammar has been avoided by
+  # pushing some extra smarts into the Lexer.
+  class Lexer
+
+    # The list of keywords passed verbatim to the parser.
+    KEYWORDS   = ["if", "else", "then", "unless", "until",
+                  "true", "false", "yes", "no", "on", "off",
+                  "and", "or", "is", "isnt", "not",
+                  "new", "return",
+                  "try", "catch", "finally", "throw",
+                  "break", "continue",
+                  "for", "in", "of", "by", "where", "while",
+                  "delete", "instanceof", "typeof",
+                  "switch", "when",
+                  "super", "extends"]
+
+    # Token matching regexes.
+    IDENTIFIER = /\A([a-zA-Z$_](\w|\$)*)/
+    NUMBER     = /\A(\b((0(x|X)[0-9a-fA-F]+)|([0-9]+(\.[0-9]+)?(e[+\-]?[0-9]+)?)))\b/i
+    STRING     = /\A(""|''|"(.*?)([^\\]|\\\\)"|'(.*?)([^\\]|\\\\)')/m
+    HEREDOC    = /\A("{6}|'{6}|"{3}\n?(.*?)\n?([ \t]*)"{3}|'{3}\n?(.*?)\n?([ \t]*)'{3})/m
+    JS         = /\A(``|`(.*?)([^\\]|\\\\)`)/m
+    OPERATOR   = /\A([+\*&|\/\-%=<>:!?]+)/
+    WHITESPACE = /\A([ \t]+)/
+    COMMENT    = /\A(((\n?[ \t]*)?#.*$)+)/
+    CODE       = /\A((-|=)>)/
+    REGEX      = /\A(\/(.*?)([^\\]|\\\\)\/[imgy]{0,4})/
+    MULTI_DENT = /\A((\n([ \t]*))+)(\.)?/
+    LAST_DENT  = /\n([ \t]*)/
+    ASSIGNMENT = /\A(:|=)\Z/
+
+    # Token cleaning regexes.
+    JS_CLEANER      = /(\A`|`\Z)/
+    MULTILINER      = /\n/
+    STRING_NEWLINES = /\n[ \t]*/
+    COMMENT_CLEANER = /(^[ \t]*#|\n[ \t]*$)/
+    NO_NEWLINE      = /\A([+\*&|\/\-%=<>:!.\\][<>=&|]*|and|or|is|isnt|not|delete|typeof|instanceof)\Z/
+    HEREDOC_INDENT  = /^[ \t]+/
+
+    # Tokens which a regular expression will never immediately follow, but which
+    # a division operator might.
+    # See: http://www.mozilla.org/js/language/js20-2002-04/rationale/syntax.html#regular-expressions
+    NOT_REGEX  = [
+      :IDENTIFIER, :NUMBER, :REGEX, :STRING,
+      ')', '++', '--', ']', '}',
+      :FALSE, :NULL, :TRUE
+    ]
+
+    # Tokens which could legitimately be invoked or indexed.
+    CALLABLE = [:IDENTIFIER, :SUPER, ')', ']', '}', :STRING]
+
+    # Scan by attempting to match tokens one character at a time. Slow and steady.
+    def tokenize(code)
+      @code    = code.chomp # Cleanup code by remove extra line breaks
+      @i       = 0          # Current character position we're parsing
+      @line    = 1          # The current line.
+      @indent  = 0          # The current indent level.
+      @indents = []         # The stack of all indent levels we are currently within.
+      @tokens  = []         # Collection of all parsed tokens in the form [:TOKEN_TYPE, value]
+      @spaced  = nil        # The last value that has a space following it.
+      while @i < @code.length
+        @chunk = @code[@i..-1]
+        extract_next_token
+      end
+      puts "original stream: #{@tokens.inspect}" if ENV['VERBOSE']
+      close_indentation
+      Rewriter.new.rewrite(@tokens)
+    end
+
+    # At every position, run through this list of attempted matches,
+    # short-circuiting if any of them succeed.
+    def extract_next_token
+      return if identifier_token
+      return if number_token
+      return if heredoc_token
+      return if string_token
+      return if js_token
+      return if regex_token
+      return if indent_token
+      return if comment_token
+      return if whitespace_token
+      return    literal_token
+    end
+
+    # Tokenizers ==========================================================
+
+    # Matches identifying literals: variables, keywords, method names, etc.
+    def identifier_token
+      return false unless identifier = @chunk[IDENTIFIER, 1]
+      # Keywords are special identifiers tagged with their own name,
+      # 'if' will result in an [:IF, "if"] token.
+      tag = KEYWORDS.include?(identifier) ? identifier.upcase.to_sym : :IDENTIFIER
+      tag = :LEADING_WHEN if tag == :WHEN && [:OUTDENT, :INDENT, "\n"].include?(last_tag)
+      @tokens[-1][0] = :PROTOTYPE_ACCESS if tag == :IDENTIFIER && last_value == '::'
+      if tag == :IDENTIFIER && last_value == '.' && !(@tokens[-2] && @tokens[-2][1] == '.')
+        if @tokens[-2][0] == "?"
+          @tokens[-1][0] = :SOAK_ACCESS
+          @tokens.delete_at(-2)
+        else
+          @tokens[-1][0] = :PROPERTY_ACCESS
+        end
+      end
+      token(tag, identifier)
+      @i += identifier.length
+    end
+
+    # Matches numbers, including decimals, hex, and exponential notation.
+    def number_token
+      return false unless number = @chunk[NUMBER, 1]
+      token(:NUMBER, number)
+      @i += number.length
+    end
+
+    # Matches strings, including multi-line strings.
+    def string_token
+      return false unless string = @chunk[STRING, 1]
+      escaped = string.gsub(STRING_NEWLINES, " \\\n")
+      token(:STRING, escaped)
+      @line += string.count("\n")
+      @i += string.length
+    end
+
+    # Matches heredocs, adjusting indentation to the correct level.
+    def heredoc_token
+      return false unless match = @chunk.match(HEREDOC)
+      doc = match[2] || match[4]
+      indent = doc.scan(HEREDOC_INDENT).min
+      doc.gsub!(/^#{indent}/, "")
+      doc.gsub!("\n", "\\n")
+      doc.gsub!('"', '\\"')
+      token(:STRING, "\"#{doc}\"")
+      @line += match[1].count("\n")
+      @i += match[1].length
+    end
+
+    # Matches interpolated JavaScript.
+    def js_token
+      return false unless script = @chunk[JS, 1]
+      token(:JS, script.gsub(JS_CLEANER, ''))
+      @i += script.length
+    end
+
+    # Matches regular expression literals.
+    def regex_token
+      return false unless regex = @chunk[REGEX, 1]
+      return false if NOT_REGEX.include?(last_tag)
+      token(:REGEX, regex)
+      @i += regex.length
+    end
+
+    # Matches and consumes comments.
+    def comment_token
+      return false unless comment = @chunk[COMMENT, 1]
+      @line += comment.scan(MULTILINER).length
+      token(:COMMENT, comment.gsub(COMMENT_CLEANER, '').split(MULTILINER))
+      token("\n", "\n")
+      @i += comment.length
+    end
+
+    # Record tokens for indentation differing from the previous line.
+    def indent_token
+      return false unless indent = @chunk[MULTI_DENT, 1]
+      @line += indent.scan(MULTILINER).size
+      @i += indent.size
+      next_character = @chunk[MULTI_DENT, 4]
+      no_newlines = next_character == '.' || (last_value.to_s.match(NO_NEWLINE) && @tokens[-2][0] != '.'  && !last_value.match(CODE))
+      return suppress_newlines(indent) if no_newlines
+      size = indent.scan(LAST_DENT).last.last.length
+      return newline_token(indent) if size == @indent
+      if size > @indent
+        token(:INDENT, size - @indent)
+        @indents << (size - @indent)
+      else
+        outdent_token(@indent - size)
+      end
+      @indent = size
+    end
+
+    # Record an oudent token or tokens, if we're moving back inwards past
+    # multiple recorded indents.
+    def outdent_token(move_out)
+      while move_out > 0 && !@indents.empty?
+        last_indent = @indents.pop
+        token(:OUTDENT, last_indent)
+        move_out -= last_indent
+      end
+      token("\n", "\n")
+    end
+
+    # Matches and consumes non-meaningful whitespace.
+    def whitespace_token
+      return false unless whitespace = @chunk[WHITESPACE, 1]
+      @spaced = last_value
+      @i += whitespace.length
+    end
+
+    # Multiple newlines get merged together.
+    # Use a trailing \ to escape newlines.
+    def newline_token(newlines)
+      token("\n", "\n") unless last_value == "\n"
+      true
+    end
+
+    # Tokens to explicitly escape newlines are removed once their job is done.
+    def suppress_newlines(newlines)
+      @tokens.pop if last_value == "\\"
+      true
+    end
+
+    # We treat all other single characters as a token. Eg.: ( ) , . !
+    # Multi-character operators are also literal tokens, so that Racc can assign
+    # the proper order of operations.
+    def literal_token
+      value = @chunk[OPERATOR, 1]
+      tag_parameters if value && value.match(CODE)
+      value ||= @chunk[0,1]
+      tag = value.match(ASSIGNMENT) ? :ASSIGN : value
+      if !@spaced.equal?(last_value) && CALLABLE.include?(last_tag)
+        tag = :CALL_START  if value == '('
+        tag = :INDEX_START if value == '['
+      end
+      token(tag, value)
+      @i += value.length
+    end
+
+    # Helpers ==========================================================
+
+    # Add a token to the results, taking note of the line number.
+    def token(tag, value)
+      @tokens << [tag, Value.new(value, @line)]
+    end
+
+    # Peek at the previous token's value.
+    def last_value
+      @tokens.last && @tokens.last[1]
+    end
+
+    # Peek at the previous token's tag.
+    def last_tag
+      @tokens.last && @tokens.last[0]
+    end
+
+    # A source of ambiguity in our grammar was parameter lists in function
+    # definitions (as opposed to argument lists in function calls). Tag
+    # parameter identifiers in order to avoid this. Also, parameter lists can
+    # make use of splats.
+    def tag_parameters
+      return if last_tag != ')'
+      i = 0
+      loop do
+        i -= 1
+        tok = @tokens[i]
+        return if !tok
+        case tok[0]
+        when :IDENTIFIER  then tok[0] = :PARAM
+        when ')'          then tok[0] = :PARAM_END
+        when '('          then return tok[0] = :PARAM_START
+        end
+      end
+    end
+
+    # Close up all remaining open blocks. IF the first token is an indent,
+    # axe it.
+    def close_indentation
+      outdent_token(@indent)
+    end
+
+  end
+end