irb 1.0.0

data/lib/irb/magic-file.rb

@@ -0,0 +1,38 @@
+# frozen_string_literal: false
+module IRB
+  class << (MagicFile = Object.new)
+    # see parser_magic_comment in parse.y
+    ENCODING_SPEC_RE = %r"coding\s*[=:]\s*([[:alnum:]\-_]+)"
+
+    def open(path)
+      io = File.open(path, 'rb')
+      line = io.gets
+      line = io.gets if line[0,2] == "#!"
+      encoding = detect_encoding(line)
+      internal_encoding = encoding
+      encoding ||= IRB.default_src_encoding
+      io.rewind
+      io.set_encoding(encoding, internal_encoding)
+
+      if block_given?
+        begin
+          return (yield io)
+        ensure
+          io.close
+        end
+      else
+        return io
+      end
+    end
+
+    private
+    def detect_encoding(line)
+      return unless line[0] == ?#
+      line = line[1..-1]
+      line = $1 if line[/-\*-\s*(.*?)\s*-*-$/]
+      return nil unless ENCODING_SPEC_RE =~ line
+      encoding = $1
+      return encoding.sub(/-(?:mac|dos|unix)/i, '')
+    end
+  end
+end

data/lib/irb/notifier.rb

@@ -0,0 +1,232 @@
+# frozen_string_literal: false
+#
+#   notifier.rb - output methods used by irb
+#   	$Release Version: 0.9.6$
+#   	$Revision$
+#   	by Keiju ISHITSUKA(keiju@ruby-lang.org)
+#
+# --
+#
+#
+#
+
+require "e2mmap"
+require_relative "output-method"
+
+module IRB
+  # An output formatter used internally by the lexer.
+  module Notifier
+    extend Exception2MessageMapper
+    def_exception :ErrUndefinedNotifier,
+      "undefined notifier level: %d is specified"
+    def_exception :ErrUnrecognizedLevel,
+      "unrecognized notifier level: %s is specified"
+
+    # Define a new Notifier output source, returning a new CompositeNotifier
+    # with the given +prefix+ and +output_method+.
+    #
+    # The optional +prefix+ will be appended to all objects being inspected
+    # during output, using the given +output_method+ as the output source. If
+    # no +output_method+ is given, StdioOutputMethod will be used, and all
+    # expressions will be sent directly to STDOUT without any additional
+    # formatting.
+    def def_notifier(prefix = "", output_method = StdioOutputMethod.new)
+      CompositeNotifier.new(prefix, output_method)
+    end
+    module_function :def_notifier
+
+    # An abstract class, or superclass, for CompositeNotifier and
+    # LeveledNotifier to inherit. It provides several wrapper methods for the
+    # OutputMethod object used by the Notifier.
+    class AbstractNotifier
+      # Creates a new Notifier object
+      def initialize(prefix, base_notifier)
+        @prefix = prefix
+        @base_notifier = base_notifier
+      end
+
+      # The +prefix+ for this Notifier, which is appended to all objects being
+      # inspected during output.
+      attr_reader :prefix
+
+      # A wrapper method used to determine whether notifications are enabled.
+      #
+      # Defaults to +true+.
+      def notify?
+        true
+      end
+
+      # See OutputMethod#print for more detail.
+      def print(*opts)
+        @base_notifier.print prefix, *opts if notify?
+      end
+
+      # See OutputMethod#printn for more detail.
+      def printn(*opts)
+        @base_notifier.printn prefix, *opts if notify?
+      end
+
+      # See OutputMethod#printf for more detail.
+      def printf(format, *opts)
+        @base_notifier.printf(prefix + format, *opts) if notify?
+      end
+
+      # See OutputMethod#puts for more detail.
+      def puts(*objs)
+        if notify?
+          @base_notifier.puts(*objs.collect{|obj| prefix + obj.to_s})
+        end
+      end
+
+      # Same as #ppx, except it uses the #prefix given during object
+      # initialization.
+      # See OutputMethod#ppx for more detail.
+      def pp(*objs)
+        if notify?
+          @base_notifier.ppx @prefix, *objs
+        end
+      end
+
+      # Same as #pp, except it concatenates the given +prefix+ with the #prefix
+      # given during object initialization.
+      #
+      # See OutputMethod#ppx for more detail.
+      def ppx(prefix, *objs)
+        if notify?
+          @base_notifier.ppx @prefix+prefix, *objs
+        end
+      end
+
+      # Execute the given block if notifications are enabled.
+      def exec_if
+        yield(@base_notifier) if notify?
+      end
+    end
+
+    # A class that can be used to create a group of notifier objects with the
+    # intent of representing a leveled notification system for irb.
+    #
+    # This class will allow you to generate other notifiers, and assign them
+    # the appropriate level for output.
+    #
+    # The Notifier class provides a class-method Notifier.def_notifier to
+    # create a new composite notifier. Using the first composite notifier
+    # object you create, sibling notifiers can be initialized with
+    # #def_notifier.
+    class CompositeNotifier < AbstractNotifier
+      # Create a new composite notifier object with the given +prefix+, and
+      # +base_notifier+ to use for output.
+      def initialize(prefix, base_notifier)
+        super
+
+        @notifiers = [D_NOMSG]
+        @level_notifier = D_NOMSG
+      end
+
+      # List of notifiers in the group
+      attr_reader :notifiers
+
+      # Creates a new LeveledNotifier in the composite #notifiers group.
+      #
+      # The given +prefix+ will be assigned to the notifier, and +level+ will
+      # be used as the index of the #notifiers Array.
+      #
+      # This method returns the newly created instance.
+      def def_notifier(level, prefix = "")
+        notifier = LeveledNotifier.new(self, level, prefix)
+        @notifiers[level] = notifier
+        notifier
+      end
+
+      # Returns the leveled notifier for this object
+      attr_reader :level_notifier
+      alias level level_notifier
+
+      # Sets the leveled notifier for this object.
+      #
+      # When the given +value+ is an instance of AbstractNotifier,
+      # #level_notifier is set to the given object.
+      #
+      # When an Integer is given, #level_notifier is set to the notifier at the
+      # index +value+ in the #notifiers Array.
+      #
+      # If no notifier exists at the index +value+ in the #notifiers Array, an
+      # ErrUndefinedNotifier exception is raised.
+      #
+      # An ErrUnrecognizedLevel exception is raised if the given +value+ is not
+      # found in the existing #notifiers Array, or an instance of
+      # AbstractNotifier
+      def level_notifier=(value)
+        case value
+        when AbstractNotifier
+          @level_notifier = value
+        when Integer
+          l = @notifiers[value]
+          Notifier.Raise ErrUndefinedNotifier, value unless l
+          @level_notifier = l
+        else
+          Notifier.Raise ErrUnrecognizedLevel, value unless l
+        end
+      end
+
+      alias level= level_notifier=
+    end
+
+    # A leveled notifier is comparable to the composite group from
+    # CompositeNotifier#notifiers.
+    class LeveledNotifier < AbstractNotifier
+      include Comparable
+
+      # Create a new leveled notifier with the given +base+, and +prefix+ to
+      # send to AbstractNotifier.new
+      #
+      # The given +level+ is used to compare other leveled notifiers in the
+      # CompositeNotifier group to determine whether or not to output
+      # notifications.
+      def initialize(base, level, prefix)
+        super(prefix, base)
+
+        @level = level
+      end
+
+      # The current level of this notifier object
+      attr_reader :level
+
+      # Compares the level of this notifier object with the given +other+
+      # notifier.
+      #
+      # See the Comparable module for more information.
+      def <=>(other)
+        @level <=> other.level
+      end
+
+      # Whether to output messages to the output method, depending on the level
+      # of this notifier object.
+      def notify?
+        @base_notifier.level >= self
+      end
+    end
+
+    # NoMsgNotifier is a LeveledNotifier that's used as the default notifier
+    # when creating a new CompositeNotifier.
+    #
+    # This notifier is used as the +zero+ index, or level +0+, for
+    # CompositeNotifier#notifiers, and will not output messages of any sort.
+    class NoMsgNotifier < LeveledNotifier
+      # Creates a new notifier that should not be used to output messages.
+      def initialize
+        @base_notifier = nil
+        @level = 0
+        @prefix = ""
+      end
+
+      # Ensures notifications are ignored, see AbstractNotifier#notify? for
+      # more information.
+      def notify?
+        false
+      end
+    end
+
+    D_NOMSG = NoMsgNotifier.new # :nodoc:
+  end
+end

data/lib/irb/output-method.rb

@@ -0,0 +1,92 @@
+# frozen_string_literal: false
+#
+#   output-method.rb - output methods used by irb
+#   	$Release Version: 0.9.6$
+#   	$Revision$
+#   	by Keiju ISHITSUKA(keiju@ruby-lang.org)
+#
+# --
+#
+#
+#
+
+require "e2mmap"
+
+module IRB
+  # An abstract output class for IO in irb. This is mainly used internally by
+  # IRB::Notifier. You can define your own output method to use with Irb.new,
+  # or Context.new
+  class OutputMethod
+    extend Exception2MessageMapper
+    def_exception :NotImplementedError, "Need to define `%s'"
+
+
+    # Open this method to implement your own output method, raises a
+    # NotImplementedError if you don't define #print in your own class.
+    def print(*opts)
+      OutputMethod.Raise NotImplementedError, "print"
+    end
+
+    # Prints the given +opts+, with a newline delimiter.
+    def printn(*opts)
+      print opts.join(" "), "\n"
+    end
+
+    # Extends IO#printf to format the given +opts+ for Kernel#sprintf using
+    # #parse_printf_format
+    def printf(format, *opts)
+      if /(%*)%I/ =~ format
+        format, opts = parse_printf_format(format, opts)
+      end
+      print sprintf(format, *opts)
+    end
+
+    # Returns an array of the given +format+ and +opts+ to be used by
+    # Kernel#sprintf, if there was a successful Regexp match in the given
+    # +format+ from #printf
+    #
+    #     %
+    #     <flag>  [#0- +]
+    #     <minimum field width> (\*|\*[1-9][0-9]*\$|[1-9][0-9]*)
+    #     <precision>.(\*|\*[1-9][0-9]*\$|[1-9][0-9]*|)?
+    #     #<length modifier>(hh|h|l|ll|L|q|j|z|t)
+    #     <conversion specifier>[diouxXeEfgGcsb%]
+    def parse_printf_format(format, opts)
+      return format, opts if $1.size % 2 == 1
+    end
+
+    # Calls #print on each element in the given +objs+, followed by a newline
+    # character.
+    def puts(*objs)
+      for obj in objs
+        print(*obj)
+        print "\n"
+      end
+    end
+
+    # Prints the given +objs+ calling Object#inspect on each.
+    #
+    # See #puts for more detail.
+    def pp(*objs)
+      puts(*objs.collect{|obj| obj.inspect})
+    end
+
+    # Prints the given +objs+ calling Object#inspect on each and appending the
+    # given +prefix+.
+    #
+    # See #puts for more detail.
+    def ppx(prefix, *objs)
+      puts(*objs.collect{|obj| prefix+obj.inspect})
+    end
+
+  end
+
+  # A standard output printer
+  class StdioOutputMethod < OutputMethod
+    # Prints the given +opts+ to standard output, see IO#print for more
+    # information.
+    def print(*opts)
+      STDOUT.print(*opts)
+    end
+  end
+end

data/lib/irb/ruby-lex.rb

@@ -0,0 +1,1180 @@
+# frozen_string_literal: false
+#
+#   irb/ruby-lex.rb - ruby lexcal analyzer
+#   	$Release Version: 0.9.6$
+#   	$Revision$
+#   	by Keiju ISHITSUKA(keiju@ruby-lang.org)
+#
+# --
+#
+#
+#
+
+require "e2mmap"
+require_relative "slex"
+require_relative "ruby-token"
+
+# :stopdoc:
+class RubyLex
+
+  extend Exception2MessageMapper
+  def_exception(:AlreadyDefinedToken, "Already defined token(%s)")
+  def_exception(:TkReading2TokenNoKey, "key nothing(key='%s')")
+  def_exception(:TkSymbol2TokenNoKey, "key nothing(key='%s')")
+  def_exception(:TkReading2TokenDuplicateError,
+                "key duplicate(token_n='%s', key='%s')")
+  def_exception(:SyntaxError, "%s")
+
+  def_exception(:TerminateLineInput, "Terminate Line Input")
+
+  include RubyToken
+
+  class << self
+    attr_accessor :debug_level
+    def debug?
+      @debug_level > 0
+    end
+  end
+  @debug_level = 0
+
+  def initialize
+    lex_init
+    set_input(STDIN)
+
+    @seek = 0
+    @exp_line_no = @line_no = 1
+    @base_char_no = 0
+    @char_no = 0
+    @rests = []
+    @readed = []
+    @here_readed = []
+
+    @indent = 0
+    @indent_stack = []
+    @lex_state = EXPR_BEG
+    @space_seen = false
+    @here_header = false
+    @post_symbeg = false
+
+    @continue = false
+    @line = ""
+
+    @skip_space = false
+    @readed_auto_clean_up = false
+    @exception_on_syntax_error = true
+
+    @prompt = nil
+  end
+
+  attr_accessor :skip_space
+  attr_accessor :readed_auto_clean_up
+  attr_accessor :exception_on_syntax_error
+
+  attr_reader :seek
+  attr_reader :char_no
+  attr_reader :line_no
+  attr_reader :indent
+
+  # io functions
+  def set_input(io, p = nil, &block)
+    @io = io
+    if p.respond_to?(:call)
+      @input = p
+    elsif block_given?
+      @input = block
+    else
+      @input = Proc.new{@io.gets}
+    end
+  end
+
+  def get_readed
+    if idx = @readed.rindex("\n")
+      @base_char_no = @readed.size - (idx + 1)
+    else
+      @base_char_no += @readed.size
+    end
+
+    readed = @readed.join("")
+    @readed = []
+    readed
+  end
+
+  def getc
+    while @rests.empty?
+      @rests.push nil unless buf_input
+    end
+    c = @rests.shift
+    if @here_header
+      @here_readed.push c
+    else
+      @readed.push c
+    end
+    @seek += 1
+    if c == "\n"
+      @line_no += 1
+      @char_no = 0
+    else
+      @char_no += 1
+    end
+    c
+  end
+
+  def gets
+    l = ""
+    while c = getc
+      l.concat(c)
+      break if c == "\n"
+    end
+    return nil if l == "" and c.nil?
+    l
+  end
+
+  def eof?
+    @io.eof?
+  end
+
+  def getc_of_rests
+    if @rests.empty?
+      nil
+    else
+      getc
+    end
+  end
+
+  def ungetc(c = nil)
+    if @here_readed.empty?
+      c2 = @readed.pop
+    else
+      c2 = @here_readed.pop
+    end
+    c = c2 unless c
+    @rests.unshift c #c =
+    @seek -= 1
+    if c == "\n"
+      @line_no -= 1
+      if idx = @readed.rindex("\n")
+        @char_no = idx + 1
+      else
+        @char_no = @base_char_no + @readed.size
+      end
+    else
+      @char_no -= 1
+    end
+  end
+
+  def peek_equal?(str)
+    chrs = str.split(//)
+    until @rests.size >= chrs.size
+      return false unless buf_input
+    end
+    @rests[0, chrs.size] == chrs
+  end
+
+  def peek_match?(regexp)
+    while @rests.empty?
+      return false unless buf_input
+    end
+    regexp =~ @rests.join("")
+  end
+
+  def peek(i = 0)
+    while @rests.size <= i
+      return nil unless buf_input
+    end
+    @rests[i]
+  end
+
+  def buf_input
+    prompt
+    line = @input.call
+    return nil unless line
+    @rests.concat line.chars.to_a
+    true
+  end
+  private :buf_input
+
+  def set_prompt(p = nil, &block)
+    p = block if block_given?
+    if p.respond_to?(:call)
+      @prompt = p
+    else
+      @prompt = Proc.new{print p}
+    end
+  end
+
+  def prompt
+    if @prompt
+      @prompt.call(@ltype, @indent, @continue, @line_no)
+    end
+  end
+
+  def initialize_input
+    @ltype = nil
+    @quoted = nil
+    @indent = 0
+    @indent_stack = []
+    @lex_state = EXPR_BEG
+    @space_seen = false
+    @here_header = false
+
+    @continue = false
+    @post_symbeg = false
+
+    prompt
+
+    @line = ""
+    @exp_line_no = @line_no
+  end
+
+  def each_top_level_statement
+    initialize_input
+    catch(:TERM_INPUT) do
+      loop do
+        begin
+          @continue = false
+          prompt
+          unless l = lex
+            throw :TERM_INPUT if @line == ''
+          else
+            @line.concat l
+            if @ltype or @continue or @indent > 0
+              next
+            end
+          end
+          if @line != "\n"
+            @line.force_encoding(@io.encoding)
+            yield @line, @exp_line_no
+          end
+          break unless l
+          @line = ''
+          @exp_line_no = @line_no
+
+          @indent = 0
+          @indent_stack = []
+          prompt
+        rescue TerminateLineInput
+          initialize_input
+          prompt
+          get_readed
+        end
+      end
+    end
+  end
+
+  def lex
+    continue = @continue
+    while tk = token
+      case tk
+      when TkNL, TkEND_OF_SCRIPT
+        @continue = continue unless continue.nil?
+        break unless @continue
+      when TkSPACE, TkCOMMENT
+      when TkSEMICOLON, TkBEGIN, TkELSE
+        @continue = continue = false
+      else
+        continue = nil
+      end
+    end
+    line = get_readed
+    if line == "" and tk.kind_of?(TkEND_OF_SCRIPT) || tk.nil?
+      nil
+    else
+      line
+    end
+  end
+
+  def token
+    @prev_seek = @seek
+    @prev_line_no = @line_no
+    @prev_char_no = @char_no
+    begin
+      begin
+        tk = @OP.match(self)
+        @space_seen = tk.kind_of?(TkSPACE)
+        @lex_state = EXPR_END if @post_symbeg && tk.kind_of?(TkOp)
+        @post_symbeg = tk.kind_of?(TkSYMBEG)
+      rescue SyntaxError
+        raise if @exception_on_syntax_error
+        tk = TkError.new(@seek, @line_no, @char_no)
+      end
+    end while @skip_space and tk.kind_of?(TkSPACE)
+    if @readed_auto_clean_up
+      get_readed
+    end
+    tk
+  end
+
+  ENINDENT_CLAUSE = [
+    "case", "class", "def", "do", "for", "if",
+    "module", "unless", "until", "while", "begin"
+  ]
+  DEINDENT_CLAUSE = ["end"
+  ]
+
+  PERCENT_LTYPE = {
+    "q" => "\'",
+    "Q" => "\"",
+    "x" => "\`",
+    "r" => "/",
+    "w" => "]",
+    "W" => "]",
+    "i" => "]",
+    "I" => "]",
+    "s" => ":"
+  }
+
+  PERCENT_PAREN = {
+    "{" => "}",
+    "[" => "]",
+    "<" => ">",
+    "(" => ")"
+  }
+
+  Ltype2Token = {
+    "\'" => TkSTRING,
+    "\"" => TkSTRING,
+    "\`" => TkXSTRING,
+    "/" => TkREGEXP,
+    "]" => TkDSTRING,
+    ":" => TkSYMBOL
+  }
+  DLtype2Token = {
+    "\"" => TkDSTRING,
+    "\`" => TkDXSTRING,
+    "/" => TkDREGEXP,
+  }
+
+  def lex_init()
+    @OP = IRB::SLex.new
+    @OP.def_rules("\0", "\004", "\032") do |op, io|
+      Token(TkEND_OF_SCRIPT)
+    end
+
+    @OP.def_rules(" ", "\t", "\f", "\r", "\13") do |op, io|
+      @space_seen = true
+      while getc =~ /[ \t\f\r\13]/; end
+      ungetc
+      Token(TkSPACE)
+    end
+
+    @OP.def_rule("#") do |op, io|
+      identify_comment
+    end
+
+    @OP.def_rule("=begin",
+                 proc{|op, io| @prev_char_no == 0 && peek(0) =~ /\s/}) do
+      |op, io|
+      @ltype = "="
+      until getc == "\n"; end
+      until peek_equal?("=end") && peek(4) =~ /\s/
+        until getc == "\n"; end
+      end
+      gets
+      @ltype = nil
+      Token(TkRD_COMMENT)
+    end
+
+    @OP.def_rule("\n") do |op, io|
+      print "\\n\n" if RubyLex.debug?
+      case @lex_state
+      when EXPR_BEG, EXPR_FNAME, EXPR_DOT
+        @continue = true
+      else
+        @continue = false
+        @lex_state = EXPR_BEG
+        until (@indent_stack.empty? ||
+            [TkLPAREN, TkLBRACK, TkLBRACE,
+             TkfLPAREN, TkfLBRACK, TkfLBRACE].include?(@indent_stack.last))
+          @indent_stack.pop
+        end
+      end
+      @here_header = false
+      @here_readed = []
+      Token(TkNL)
+    end
+
+    @OP.def_rules("*", "**",
+                  "=", "==", "===",
+                  "=~", "<=>",
+                  "<", "<=",
+                  ">", ">=", ">>",
+                  "!", "!=", "!~") do
+      |op, io|
+      case @lex_state
+      when EXPR_FNAME, EXPR_DOT
+        @lex_state = EXPR_ARG
+      else
+        @lex_state = EXPR_BEG
+      end
+      Token(op)
+    end
+
+    @OP.def_rules("<<") do
+      |op, io|
+      tk = nil
+      if @lex_state != EXPR_END && @lex_state != EXPR_CLASS &&
+          (@lex_state != EXPR_ARG || @space_seen)
+        c = peek(0)
+        if /[-~"'`\w]/ =~ c
+          tk = identify_here_document
+        end
+      end
+      unless tk
+        tk = Token(op)
+        case @lex_state
+        when EXPR_FNAME, EXPR_DOT
+          @lex_state = EXPR_ARG
+        else
+          @lex_state = EXPR_BEG
+        end
+      end
+      tk
+    end
+
+    @OP.def_rules("'", '"') do
+      |op, io|
+      identify_string(op)
+    end
+
+    @OP.def_rules("`") do
+      |op, io|
+      if @lex_state == EXPR_FNAME
+        @lex_state = EXPR_END
+        Token(op)
+      else
+        identify_string(op)
+      end
+    end
+
+    @OP.def_rules('?') do
+      |op, io|
+      if @lex_state == EXPR_END
+        @lex_state = EXPR_BEG
+        Token(TkQUESTION)
+      else
+        ch = getc
+        if @lex_state == EXPR_ARG && ch =~ /\s/
+          ungetc
+          @lex_state = EXPR_BEG;
+          Token(TkQUESTION)
+        else
+          if (ch == '\\')
+            read_escape
+          end
+          @lex_state = EXPR_END
+          Token(TkINTEGER)
+        end
+      end
+    end
+
+    @OP.def_rules("&", "&&", "|", "||") do
+      |op, io|
+      @lex_state = EXPR_BEG
+      Token(op)
+    end
+
+    @OP.def_rules("+=", "-=", "*=", "**=",
+                  "&=", "|=", "^=", "<<=", ">>=", "||=", "&&=") do
+      |op, io|
+      @lex_state = EXPR_BEG
+      op =~ /^(.*)=$/
+      Token(TkOPASGN, $1)
+    end
+
+    @OP.def_rule("+@", proc{|op, io| @lex_state == EXPR_FNAME}) do
+      |op, io|
+      @lex_state = EXPR_ARG
+      Token(op)
+    end
+
+    @OP.def_rule("-@", proc{|op, io| @lex_state == EXPR_FNAME}) do
+      |op, io|
+      @lex_state = EXPR_ARG
+      Token(op)
+    end
+
+    @OP.def_rules("+", "-") do
+      |op, io|
+      catch(:RET) do
+        if @lex_state == EXPR_ARG
+          if @space_seen and peek(0) =~ /[0-9]/
+            throw :RET, identify_number
+          else
+            @lex_state = EXPR_BEG
+          end
+        elsif @lex_state != EXPR_END and peek(0) =~ /[0-9]/
+          throw :RET, identify_number
+        else
+          @lex_state = EXPR_BEG
+        end
+        Token(op)
+      end
+    end
+
+    @OP.def_rule(".") do
+      |op, io|
+      @lex_state = EXPR_BEG
+      if peek(0) =~ /[0-9]/
+        ungetc
+        identify_number
+      else
+        # for "obj.if" etc.
+        @lex_state = EXPR_DOT
+        Token(TkDOT)
+      end
+    end
+
+    @OP.def_rules("..", "...") do
+      |op, io|
+      @lex_state = EXPR_BEG
+      Token(op)
+    end
+
+    lex_int2
+  end
+
+  def lex_int2
+    @OP.def_rules("]", "}", ")") do
+      |op, io|
+      @lex_state = EXPR_END
+      @indent -= 1
+      @indent_stack.pop
+      Token(op)
+    end
+
+    @OP.def_rule(":") do
+      |op, io|
+      if @lex_state == EXPR_END || peek(0) =~ /\s/
+        @lex_state = EXPR_BEG
+        Token(TkCOLON)
+      else
+        @lex_state = EXPR_FNAME
+        Token(TkSYMBEG)
+      end
+    end
+
+    @OP.def_rule("::") do
+       |op, io|
+      if @lex_state == EXPR_BEG or @lex_state == EXPR_ARG && @space_seen
+        @lex_state = EXPR_BEG
+        Token(TkCOLON3)
+      else
+        @lex_state = EXPR_DOT
+        Token(TkCOLON2)
+      end
+    end
+
+    @OP.def_rule("/") do
+      |op, io|
+      if @lex_state == EXPR_BEG || @lex_state == EXPR_MID
+        identify_string(op)
+      elsif peek(0) == '='
+        getc
+        @lex_state = EXPR_BEG
+        Token(TkOPASGN, "/") #/)
+      elsif @lex_state == EXPR_ARG and @space_seen and peek(0) !~ /\s/
+        identify_string(op)
+      else
+        @lex_state = EXPR_BEG
+        Token("/") #/)
+      end
+    end
+
+    @OP.def_rules("^") do
+      |op, io|
+      @lex_state = EXPR_BEG
+      Token("^")
+    end
+
+    @OP.def_rules(",") do
+      |op, io|
+      @lex_state = EXPR_BEG
+      Token(op)
+    end
+
+    @OP.def_rules(";") do
+      |op, io|
+      @lex_state = EXPR_BEG
+      until (@indent_stack.empty? ||
+          [TkLPAREN, TkLBRACK, TkLBRACE,
+           TkfLPAREN, TkfLBRACK, TkfLBRACE].include?(@indent_stack.last))
+        @indent_stack.pop
+      end
+      Token(op)
+    end
+
+    @OP.def_rule("~") do
+      |op, io|
+      @lex_state = EXPR_BEG
+      Token("~")
+    end
+
+    @OP.def_rule("~@", proc{|op, io| @lex_state == EXPR_FNAME}) do
+      |op, io|
+      @lex_state = EXPR_BEG
+      Token("~")
+    end
+
+    @OP.def_rule("(") do
+      |op, io|
+      @indent += 1
+      if @lex_state == EXPR_BEG || @lex_state == EXPR_MID
+        @lex_state = EXPR_BEG
+        tk_c = TkfLPAREN
+      else
+        @lex_state = EXPR_BEG
+        tk_c = TkLPAREN
+      end
+      @indent_stack.push tk_c
+      Token(tk_c)
+    end
+
+    @OP.def_rule("[]", proc{|op, io| @lex_state == EXPR_FNAME}) do
+      |op, io|
+      @lex_state = EXPR_ARG
+      Token("[]")
+    end
+
+    @OP.def_rule("[]=", proc{|op, io| @lex_state == EXPR_FNAME}) do
+      |op, io|
+      @lex_state = EXPR_ARG
+      Token("[]=")
+    end
+
+    @OP.def_rule("[") do
+      |op, io|
+      @indent += 1
+      if @lex_state == EXPR_FNAME
+        tk_c = TkfLBRACK
+      else
+        if @lex_state == EXPR_BEG || @lex_state == EXPR_MID
+          tk_c = TkLBRACK
+        elsif @lex_state == EXPR_ARG && @space_seen
+          tk_c = TkLBRACK
+        else
+          tk_c = TkfLBRACK
+        end
+        @lex_state = EXPR_BEG
+      end
+      @indent_stack.push tk_c
+      Token(tk_c)
+    end
+
+    @OP.def_rule("{") do
+      |op, io|
+      @indent += 1
+      if @lex_state != EXPR_END && @lex_state != EXPR_ARG
+        tk_c = TkLBRACE
+      else
+        tk_c = TkfLBRACE
+      end
+      @lex_state = EXPR_BEG
+      @indent_stack.push tk_c
+      Token(tk_c)
+    end
+
+    @OP.def_rule('\\') do
+      |op, io|
+      if getc == "\n"
+        @space_seen = true
+        @continue = true
+        Token(TkSPACE)
+      else
+        read_escape
+        Token("\\")
+      end
+    end
+
+    @OP.def_rule('%') do
+      |op, io|
+      if @lex_state == EXPR_BEG || @lex_state == EXPR_MID
+        identify_quotation
+      elsif peek(0) == '='
+        getc
+        Token(TkOPASGN, :%)
+      elsif @lex_state == EXPR_ARG and @space_seen and peek(0) !~ /\s/
+        identify_quotation
+      else
+        @lex_state = EXPR_BEG
+        Token("%") #))
+      end
+    end
+
+    @OP.def_rule('$') do
+      |op, io|
+      identify_gvar
+    end
+
+    @OP.def_rule('@') do
+      |op, io|
+      if peek(0) =~ /[\w@]/
+        ungetc
+        identify_identifier
+      else
+        Token("@")
+      end
+    end
+
+    @OP.def_rule("") do
+      |op, io|
+      printf "MATCH: start %s: %s\n", op, io.inspect if RubyLex.debug?
+      if peek(0) =~ /[0-9]/
+        t = identify_number
+      elsif peek(0) =~ /[^\x00-\/:-@\[-^`{-\x7F]/
+        t = identify_identifier
+      end
+      printf "MATCH: end %s: %s\n", op, io.inspect if RubyLex.debug?
+      t
+    end
+
+    p @OP if RubyLex.debug?
+  end
+
+  def identify_gvar
+    @lex_state = EXPR_END
+
+    case ch = getc
+    when /[~_*$?!@\/\\;,=:<>".]/   #"
+      Token(TkGVAR, "$" + ch)
+    when "-"
+      Token(TkGVAR, "$-" + getc)
+    when "&", "`", "'", "+"
+      Token(TkBACK_REF, "$"+ch)
+    when /[1-9]/
+      while getc =~ /[0-9]/; end
+      ungetc
+      Token(TkNTH_REF)
+    when /\w/
+      ungetc
+      ungetc
+      identify_identifier
+    else
+      ungetc
+      Token("$")
+    end
+  end
+
+  def identify_identifier
+    token = ""
+    if peek(0) =~ /[$@]/
+      token.concat(c = getc)
+      if c == "@" and peek(0) == "@"
+        token.concat getc
+      end
+    end
+
+    while (ch = getc) =~ /[^\x00-\/:-@\[-^`{-\x7F]/
+      print ":", ch, ":" if RubyLex.debug?
+      token.concat ch
+    end
+    ungetc
+
+    if (ch == "!" || ch == "?") && token[0,1] =~ /\w/ && peek(0) != "="
+      token.concat getc
+    end
+
+    # almost fix token
+
+    case token
+    when /^\$/
+      return Token(TkGVAR, token)
+    when /^\@\@/
+      @lex_state = EXPR_END
+      # p Token(TkCVAR, token)
+      return Token(TkCVAR, token)
+    when /^\@/
+      @lex_state = EXPR_END
+      return Token(TkIVAR, token)
+    end
+
+    if @lex_state != EXPR_DOT
+      print token, "\n" if RubyLex.debug?
+
+      token_c, *trans = TkReading2Token[token]
+      if token_c
+        # reserved word?
+
+        if (@lex_state != EXPR_BEG &&
+            @lex_state != EXPR_FNAME &&
+            trans[1])
+          # modifiers
+          token_c = TkSymbol2Token[trans[1]]
+          @lex_state = trans[0]
+        else
+          if @lex_state != EXPR_FNAME and peek(0) != ':'
+            if ENINDENT_CLAUSE.include?(token)
+              # check for ``class = val'' etc.
+              valid = true
+              case token
+              when "class"
+                valid = false unless peek_match?(/^\s*(<<|\w|::)/)
+              when "def"
+                valid = false if peek_match?(/^\s*(([+\-\/*&\|^]|<<|>>|\|\||\&\&)=|\&\&|\|\|)/)
+              when "do"
+                valid = false if peek_match?(/^\s*([+\-\/*]?=|\*|<|>|\&)/)
+              when *ENINDENT_CLAUSE
+                valid = false if peek_match?(/^\s*([+\-\/*]?=|\*|<|>|\&|\|)/)
+              else
+                # no nothing
+              end
+              if valid
+                if token == "do"
+                  if ![TkFOR, TkWHILE, TkUNTIL].include?(@indent_stack.last)
+                    @indent += 1
+                    @indent_stack.push token_c
+                  end
+                else
+                  @indent += 1
+                  @indent_stack.push token_c
+                end
+              end
+
+            elsif DEINDENT_CLAUSE.include?(token)
+              @indent -= 1
+              @indent_stack.pop
+            end
+            @lex_state = trans[0]
+          else
+            @lex_state = EXPR_END
+          end
+        end
+        return Token(token_c, token)
+      end
+    end
+
+    if @lex_state == EXPR_FNAME
+      @lex_state = EXPR_END
+      if peek(0) == '='
+        token.concat getc
+      end
+    elsif @lex_state == EXPR_BEG || @lex_state == EXPR_DOT
+      @lex_state = EXPR_ARG
+    else
+      @lex_state = EXPR_END
+    end
+
+    if token[0, 1] =~ /[A-Z]/
+      return Token(TkCONSTANT, token)
+    elsif token[token.size - 1, 1] =~ /[!?]/
+      return Token(TkFID, token)
+    else
+      return Token(TkIDENTIFIER, token)
+    end
+  end
+
+  def identify_here_document
+    ch = getc
+    if ch == "-" || ch == "~"
+      ch = getc
+      indent = true
+    end
+    if /['"`]/ =~ ch
+      lt = ch
+      quoted = ""
+      while (c = getc) && c != lt
+        quoted.concat c
+      end
+    else
+      lt = '"'
+      quoted = ch.dup
+      while (c = getc) && c =~ /\w/
+        quoted.concat c
+      end
+      ungetc
+    end
+
+    ltback, @ltype = @ltype, lt
+    reserve = []
+    while ch = getc
+      reserve.push ch
+      if ch == "\\"
+        reserve.push ch = getc
+      elsif ch == "\n"
+        break
+      end
+    end
+
+    @here_header = false
+
+    line = ""
+    while ch = getc
+      if ch == "\n"
+        if line == quoted
+          break
+        end
+        line = ""
+      else
+        line.concat ch unless indent && line == "" && /\s/ =~ ch
+        if @ltype != "'" && ch == "#" && peek(0) == "{"
+          identify_string_dvar
+        end
+      end
+    end
+
+    @here_header = true
+    @here_readed.concat reserve
+    while ch = reserve.pop
+      ungetc ch
+    end
+
+    @ltype = ltback
+    @lex_state = EXPR_END
+    Token(Ltype2Token[lt])
+  end
+
+  def identify_quotation
+    ch = getc
+    if lt = PERCENT_LTYPE[ch]
+      ch = getc
+    elsif ch =~ /\W/
+      lt = "\""
+    else
+      RubyLex.fail SyntaxError, "unknown type of %string"
+    end
+    @quoted = ch unless @quoted = PERCENT_PAREN[ch]
+    identify_string(lt, @quoted)
+  end
+
+  def identify_number
+    @lex_state = EXPR_END
+
+    if peek(0) == "0" && peek(1) !~ /[.eE]/
+      getc
+      case peek(0)
+      when /[xX]/
+        ch = getc
+        match = /[0-9a-fA-F_]/
+      when /[bB]/
+        ch = getc
+        match = /[01_]/
+      when /[oO]/
+        ch = getc
+        match = /[0-7_]/
+      when /[dD]/
+        ch = getc
+        match = /[0-9_]/
+      when /[0-7]/
+        match = /[0-7_]/
+      when /[89]/
+        RubyLex.fail SyntaxError, "Invalid octal digit"
+      else
+        return Token(TkINTEGER)
+      end
+
+      len0 = true
+      non_digit = false
+      while ch = getc
+        if match =~ ch
+          if ch == "_"
+            if non_digit
+              RubyLex.fail SyntaxError, "trailing `#{ch}' in number"
+            else
+              non_digit = ch
+            end
+          else
+            non_digit = false
+            len0 = false
+          end
+        else
+          ungetc
+          if len0
+            RubyLex.fail SyntaxError, "numeric literal without digits"
+          end
+          if non_digit
+            RubyLex.fail SyntaxError, "trailing `#{non_digit}' in number"
+          end
+          break
+        end
+      end
+      return Token(TkINTEGER)
+    end
+
+    type = TkINTEGER
+    allow_point = true
+    allow_e = true
+    non_digit = false
+    while ch = getc
+      case ch
+      when /[0-9]/
+        non_digit = false
+      when "_"
+        non_digit = ch
+      when allow_point && "."
+        if non_digit
+          RubyLex.fail SyntaxError, "trailing `#{non_digit}' in number"
+        end
+        type = TkFLOAT
+        if peek(0) !~ /[0-9]/
+          type = TkINTEGER
+          ungetc
+          break
+        end
+        allow_point = false
+      when allow_e && "e", allow_e && "E"
+        if non_digit
+          RubyLex.fail SyntaxError, "trailing `#{non_digit}' in number"
+        end
+        type = TkFLOAT
+        if peek(0) =~ /[+-]/
+          getc
+        end
+        allow_e = false
+        allow_point = false
+        non_digit = ch
+      else
+        if non_digit
+          RubyLex.fail SyntaxError, "trailing `#{non_digit}' in number"
+        end
+        ungetc
+        break
+      end
+    end
+    Token(type)
+  end
+
+  def identify_string(ltype, quoted = ltype)
+    @ltype = ltype
+    @quoted = quoted
+    subtype = nil
+    begin
+      nest = 0
+      while ch = getc
+        if @quoted == ch and nest == 0
+          break
+        elsif @ltype != "'" && ch == "#" && peek(0) == "{"
+          identify_string_dvar
+        elsif @ltype != "'" && @ltype != "]" && @ltype != ":" and ch == "#"
+          subtype = true
+        elsif ch == '\\' and @ltype == "'" #'
+          case ch = getc
+          when "\\", "\n", "'"
+          else
+            ungetc
+          end
+        elsif ch == '\\' #'
+          read_escape
+        end
+        if PERCENT_PAREN.values.include?(@quoted)
+          if PERCENT_PAREN[ch] == @quoted
+            nest += 1
+          elsif ch == @quoted
+            nest -= 1
+          end
+        end
+      end
+      if @ltype == "/"
+        while /[imxoesun]/ =~ peek(0)
+          getc
+        end
+      end
+      if subtype
+        Token(DLtype2Token[ltype])
+      else
+        Token(Ltype2Token[ltype])
+      end
+    ensure
+      @ltype = nil
+      @quoted = nil
+      @lex_state = EXPR_END
+    end
+  end
+
+  def identify_string_dvar
+    begin
+      getc
+
+      reserve_continue = @continue
+      reserve_ltype = @ltype
+      reserve_indent = @indent
+      reserve_indent_stack = @indent_stack
+      reserve_state = @lex_state
+      reserve_quoted = @quoted
+
+      @ltype = nil
+      @quoted = nil
+      @indent = 0
+      @indent_stack = []
+      @lex_state = EXPR_BEG
+
+      loop do
+        @continue = false
+        prompt
+        tk = token
+        if @ltype or @continue or @indent >= 0
+          next
+        end
+        break if tk.kind_of?(TkRBRACE)
+      end
+    ensure
+      @continue = reserve_continue
+      @ltype = reserve_ltype
+      @indent = reserve_indent
+      @indent_stack = reserve_indent_stack
+      @lex_state = reserve_state
+      @quoted = reserve_quoted
+    end
+  end
+
+  def identify_comment
+    @ltype = "#"
+
+    while ch = getc
+      if ch == "\n"
+        @ltype = nil
+        ungetc
+        break
+      end
+    end
+    return Token(TkCOMMENT)
+  end
+
+  def read_escape
+    case ch = getc
+    when "\n", "\r", "\f"
+    when "\\", "n", "t", "r", "f", "v", "a", "e", "b", "s" #"
+    when /[0-7]/
+      ungetc ch
+      3.times do
+        case ch = getc
+        when /[0-7]/
+        when nil
+          break
+        else
+          ungetc
+          break
+        end
+      end
+
+    when "x"
+      2.times do
+        case ch = getc
+        when /[0-9a-fA-F]/
+        when nil
+          break
+        else
+          ungetc
+          break
+        end
+      end
+
+    when "M"
+      if (ch = getc) != '-'
+        ungetc
+      else
+        if (ch = getc) == "\\" #"
+          read_escape
+        end
+      end
+
+    when "C", "c" #, "^"
+      if ch == "C" and (ch = getc) != "-"
+        ungetc
+      elsif (ch = getc) == "\\" #"
+        read_escape
+      end
+    else
+      # other characters
+    end
+  end
+end
+# :startdoc: