ruby_parser 3.13.1 → 3.21.0

data/lib/ruby_parser_extras.rb

@@ -1,4 +1,6 @@
 # encoding: ASCII-8BIT
+# frozen_string_literal: true
+# TODO: remove encoding comment
 
 require "sexp"
 require "ruby_lexer"
@@ -6,12 +8,49 @@ require "timeout"
 require "rp_extensions"
 require "rp_stringscanner"
 
+class Sexp
+  def check_line_numbers
+    raise "bad nil line for:\n%s" % [self.pretty_inspect] if nil_line?
+    raise "bad line number for:\n%s" % [self.pretty_inspect] unless
+      Integer === self.line &&
+      self.line >= 1 &&
+      self.line <= self.line_min
+  end
+
+  ##
+  # Returns the minimum line number of the children of self.
+
+  def line_min
+    @line_min ||= [self.deep_each.map(&:line).min, self.line].compact.min
+  end
+
+  def nil_line?
+    self.deep_each.map(&:line).any?(&:nil?)
+  end
+end
+
 module RubyParserStuff
-  VERSION = "3.13.1"
+  VERSION = "3.21.0"
 
-  attr_accessor :lexer, :in_def, :in_single, :file
+  attr_accessor :lexer, :in_def, :in_single, :file, :in_argdef
   attr_accessor :in_kwarg
-  attr_reader :env, :comments
+  attr_reader :env
+
+  ##
+  # Canonicalize conditionals. Eg:
+  #
+  #   not x ? a : b
+  #
+  # becomes:
+  #
+  #   x ? b : a
+
+  attr_accessor :canonicalize_conditions
+
+  ##
+  # The last token type returned from #next_token
+
+  attr_accessor :last_token_type
 
   $good20 = []
 
@@ -31,6 +70,29 @@ module RubyParserStuff
     end
   end
 
+  ##
+  # for pure ruby systems only
+
+  def do_parse
+    _racc_do_parse_rb(_racc_setup, false)
+  end if ENV["PURE_RUBY"] || ENV["CHECK_LINE_NUMS"]
+
+  if ENV["CHECK_LINE_NUMS"] then
+    def _racc_do_reduce arg, act
+      x = super
+
+      @racc_vstack.grep(Sexp).each do |sexp|
+        sexp.check_line_numbers
+      end
+      x
+    end
+  end
+
+  ARG_TYPES = [:arglist, :call_args, :array, :args].map { |k|
+    [k, true]
+  }.to_h
+
+  # TODO: remove
   has_enc = "".respond_to? :encoding
 
   # This is in sorted order of occurrence according to
@@ -48,91 +110,79 @@ module RubyParserStuff
     Encoding::EUC_JP
   ] if has_enc
 
-  def syntax_error msg
-    raise RubyParser::SyntaxError, msg
-  end
+  JUMP_TYPE = [:return, :next, :break, :yield].map { |k| [k, true] }.to_h
 
-  def arg_blk_pass node1, node2 # TODO: nuke
-    node1 = s(:arglist, node1) unless [:arglist, :call_args, :array, :args].include? node1.sexp_type 
-    node1 << node2 if node2
-    node1
+  TAB_WIDTH = 8
+
+  def initialize(options = {})
+    super()
+
+    v = self.class.name[/[23]\d/]
+    raise "Bad Class name #{self.class}" unless v
+
+    self.lexer = RubyLexer.new v && v.to_i
+    self.lexer.parser = self
+    self.in_kwarg = false
+    self.in_argdef = false
+
+    @env = RubyParserStuff::Environment.new
+
+    @canonicalize_conditions = true
+
+    self.reset
   end
 
   def arg_concat node1, node2 # TODO: nuke
     raise "huh" unless node2
-    node1 << s(:splat, node2).compact
-    node1
-  end
 
-  def clean_mlhs sexp
-    case sexp.sexp_type
-    when :masgn then
-      if sexp.size == 2 and sexp[1].sexp_type == :array then
-        s(:masgn, *sexp[1].sexp_body.map { |sub| clean_mlhs sub })
-      else
-        debug20 5
-        sexp
-      end
-    when :gasgn, :iasgn, :lasgn, :cvasgn then
-      if sexp.size == 2 then
-        sexp.last
-      else
-        debug20 7
-        sexp # optional value
-      end
-    else
-      raise "unsupported type: #{sexp.inspect}"
-    end
+    splat = s(:splat, node2)
+    splat.line node2.line
+
+    node1 << splat
   end
 
-  def block_var *args
-    result = self.args args
-    result.sexp_type = :masgn
-    result
+  def argl x
+    x = s(:arglist, x) if x and x.sexp_type == :array
+    x
   end
 
-  def array_to_hash array
-    case array.sexp_type
-    when :kwsplat then
-      array
+  def args args
+    result = s(:args)
+
+    ss = args.grep Sexp
+    if ss.empty? then
+      result.line lexer.lineno
     else
-      s(:hash, *array.sexp_body)
+      result.line ss.first.line
+      result.line_max = ss.first.line_max
     end
-  end
-
-  def call_args args
-    result = s(:call_args)
 
     args.each do |arg|
-      case arg
-      when Sexp then
-        case arg.sexp_type
-        when :array, :args, :call_args then # HACK? remove array at some point
-          result.concat arg.sexp_body
-        else
-          result << arg
-        end
-      when Symbol then
-        result << arg
-      when ",", nil then
-        # ignore
-      else
-        raise "unhandled: #{arg.inspect} in #{args.inspect}"
+      if arg.instance_of? Array and arg.size == 2 and arg.last.is_a? Numeric then
+        arg = arg.first
       end
-    end
 
-    result
-  end
-
-  def args args
-    result = s(:args)
-
-    args.each do |arg|
       case arg
       when Sexp then
         case arg.sexp_type
         when :args, :block, :array, :call_args then # HACK call_args mismatch
-          result.concat arg.sexp_body
+          rest = arg.sexp_body
+
+          rest.map! { |x|
+            if x.instance_of? Array and x.size == 2 and Numeric === x.last then
+              x.first
+            else
+              x
+            end
+          }
+
+          result.concat rest
+        when :forward_args then
+          self.env[:*]  = :lvar # TODO: arg_var(p, idFWD_REST) ?
+          self.env[:**] = :lvar
+          self.env[:&]  = :lvar
+
+          result << arg
         when :block_arg then
           result << :"&#{arg.last}"
         when :shadow then
@@ -152,6 +202,8 @@ module RubyParserStuff
         name = arg.to_s.delete("&*")
         self.env[name.to_sym] = :lvar unless name.empty?
         result << arg
+      when true, false then
+        self.in_kwarg = arg
       when ",", "|", ";", "(", ")", nil then
         # ignore
       else
@@ -162,21 +214,50 @@ module RubyParserStuff
     result
   end
 
+  def end_args args
+    lexer.lex_state     = RubyLexer::State::Values::EXPR_BEG
+    lexer.command_start = true
+    self.args args
+  end
+
+  def attrset_id? id
+    id =~ /^\[\]=$|^\w+=$/
+  end
+
+  def endless_method_name defn_or_defs
+    name = defn_or_defs[1]
+    name = defn_or_defs[2] unless Symbol === name
+
+    if attrset_id? name then
+      yyerror "setter method cannot be defined in an endless method definition"
+    end
+
+    # TODO? token_info_drop(p, "def", loc->beg_pos);
+  end
+
+  def array_to_hash array
+    case array.sexp_type
+    when :kwsplat then
+      array
+    else
+      s(:hash, *array.sexp_body).line array.line
+    end
+  end
+
   def aryset receiver, index
     index ||= s()
-    s(:attrasgn, receiver, :"[]=", *index.sexp_body).compact # [].sexp_body => nil
+    l = receiver.line
+    result = s(:attrasgn, receiver, :"[]=",
+               *index.sexp_body).compact # [].sexp_body => nil
+    result.line = l
+    result
   end
 
   def assignable(lhs, value = nil)
-    id = lhs.to_sym unless Sexp === lhs
-    id = id.to_sym if Sexp === id
-
-    raise "write a test 1" if id.to_s =~ /^(?:self|nil|true|false|__LINE__|__FILE__)$/
-
-    raise SyntaxError, "Can't change the value of #{id}" if
-      id.to_s =~ /^(?:self|nil|true|false|__LINE__|__FILE__)$/
+    id, line = lhs
+    id = id.to_sym
 
-    result = case id.to_s
+    result = case id
              when /^@@/ then
                asgn = in_def || in_single > 0
                s((asgn ? :cvasgn : :cvdecl), id)
@@ -198,8 +279,22 @@ module RubyParserStuff
     self.env[id] ||= :lvar if result.sexp_type == :lasgn
 
     result << value if value
+    result.line line
+    result
+  end
 
-    return result
+  def backref_assign_error ref
+    # TODO: need a test for this... obviously
+    case ref.sexp_type
+    when :nth_ref then
+      raise "write a test 2"
+      raise SyntaxError, "Can't set variable %p" % ref.last
+    when :back_ref then
+      raise "write a test 3"
+      raise SyntaxError, "Can't set back reference %p" % ref.last
+    else
+      raise "Unknown backref type: #{ref.inspect}"
+    end
   end
 
   def block_append(head, tail)
@@ -209,12 +304,81 @@ module RubyParserStuff
     line = [head.line, tail.line].compact.min
 
     head = remove_begin(head)
-    head = s(:block, head) unless head.node_type == :block
+    head = s(:block, head).line(line) unless head.sexp_type == :block
 
-    head.line = line
+    # head.line = line
     head << tail
   end
 
+  def block_dup_check call_or_args, block
+    syntax_error "Both block arg and actual block given." if
+      block and call_or_args.block_pass?
+  end
+
+  def block_var *args
+    result = self.args args
+    result.sexp_type = :masgn
+    result
+  end
+
+  def call_args args
+    result = s(:call_args)
+
+    a = args.grep(Sexp).first
+    if a then
+      result.line a.line
+    else
+      result.line lexer.lineno
+    end
+
+    args.each do |arg|
+      # ruby 3.0+ TODO: next if arg in [String, Integer] # eg ["(", 1]
+      next if arg.class == Array && arg.map(&:class) == [String, Integer]
+
+      case arg
+      when Sexp then
+        case arg.sexp_type
+        when :array, :args, :call_args then # HACK? remove array at some point
+          result.concat arg.sexp_body
+        else
+          result << arg
+        end
+      when Symbol then
+        result << arg
+      when Array then
+        id, _line = arg
+        result << id
+      when ",", nil, "(" then
+        # ignore
+      else
+        raise "unhandled: #{arg.inspect} in #{args.inspect}"
+      end
+    end
+
+    result
+  end
+
+  def clean_mlhs sexp
+    case sexp.sexp_type
+    when :masgn then
+      if sexp.size == 2 and sexp[1].sexp_type == :array then
+        s(:masgn, *sexp[1].sexp_body.map { |sub| clean_mlhs sub })
+      else
+        debug20 5
+        sexp
+      end
+    when :gasgn, :iasgn, :lasgn, :cvasgn then
+      if sexp.size == 2 then
+        sexp.last
+      else
+        debug20 7
+        sexp # optional value
+      end
+    else
+      raise "unsupported type: #{sexp.inspect}"
+    end
+  end
+
   def cond node
     return nil if node.nil?
     node = value_expr node
@@ -222,60 +386,116 @@ module RubyParserStuff
     case node.sexp_type
     when :lit then
       if Regexp === node.last then
-        return s(:match, node)
+        s(:match, node)
       else
-        return node
+        node
       end
     when :and then
-      return s(:and, cond(node[1]), cond(node[2]))
+      _, lhs, rhs = node
+      s(:and,  cond(lhs), cond(rhs))
     when :or then
-      return s(:or,  cond(node[1]), cond(node[2]))
+      _, lhs, rhs = node
+      s(:or,  cond(lhs), cond(rhs))
     when :dot2 then
       label = "flip#{node.hash}"
       env[label] = :lvar
       _, lhs, rhs = node
-      return s(:flip2, lhs, rhs)
+      s(:flip2, lhs, rhs) # TODO: recurse?
     when :dot3 then
       label = "flip#{node.hash}"
       env[label] = :lvar
       _, lhs, rhs = node
-      return s(:flip3, lhs, rhs)
+      s(:flip3, lhs, rhs)
     else
-      return node
-    end
+      node
+    end.line node.line
   end
 
-  ##
-  # for pure ruby systems only
+  def dedent sexp
+    dedent_count = dedent_size sexp
 
-  def do_parse
-    _racc_do_parse_rb(_racc_setup, false)
-  end if ENV['PURE_RUBY']
+    skip_one = false
+    sexp.map { |obj|
+      case obj
+      when Symbol then
+        obj
+      when String then
+        obj.lines.map { |l| remove_whitespace_width l, dedent_count }.join
+      when Sexp then
+        case obj.sexp_type
+        when :evstr then
+          skip_one = true
+          obj
+        when :str then
+          _, str = obj
+          str = if skip_one then
+                  skip_one = false
+                  s1, *rest = str.lines
+                  s1 + rest.map { |l| remove_whitespace_width l, dedent_count }.join
+                else
+                  str.lines.map { |l| remove_whitespace_width l, dedent_count }.join
+                end
 
-  def new_match lhs, rhs
-    if lhs then
-      case lhs.sexp_type
-      when :dregx, :dregx_once then
-        return s(:match2, lhs, rhs).line(lhs.line)
-      when :lit then
-        return s(:match2, lhs, rhs).line(lhs.line) if Regexp === lhs.last
+          s(:str, str).line obj.line
+        else
+          warn "unprocessed sexp %p" % [obj]
+        end
+      else
+        warn "unprocessed: %p" % [obj]
       end
-    end
+    }
+  end
 
-    if rhs then
-      case rhs.sexp_type
-      when :dregx, :dregx_once then
-        return s(:match3, rhs, lhs).line(lhs.line)
-      when :lit then
-        return s(:match3, rhs, lhs).line(lhs.line) if Regexp === rhs.last
+  def dedent_size sexp
+    skip_one = false
+    sexp.flat_map { |s|
+      case s
+      when Symbol then
+        next
+      when String then
+        s.lines
+      when Sexp then
+        case s.sexp_type
+        when :evstr then
+          skip_one = true
+          next
+        when :str then
+          _, str = s
+          lines = str.lines
+          if skip_one then
+            skip_one = false
+            lines.shift
+          end
+          lines
+        else
+          warn "unprocessed sexp %p" % [s]
+        end
+      else
+        warn "unprocessed: %p" % [s]
+      end.map { |l| whitespace_width l }
+    }.compact.min
+  end
+
+  def dedent_string string, width
+    characters_skipped = 0
+    indentation_skipped = 0
+
+    string.chars.each do |char|
+      break if indentation_skipped >= width
+      if char == " "
+        characters_skipped += 1
+        indentation_skipped += 1
+      elsif char == "\t"
+        proposed = TAB_WIDTH * (indentation_skipped / TAB_WIDTH + 1)
+        break if proposed > width
+        characters_skipped += 1
+        indentation_skipped = proposed
       end
     end
-
-    return new_call(lhs, :"=~", argl(rhs)).line(lhs.line)
+    string[characters_skipped..-1]
   end
 
   def gettable(id)
-    lineno = id.lineno if id.respond_to? :lineno
     id = id.to_sym if String === id
 
     result = case id.to_s
@@ -296,40 +516,99 @@ module RubyParserStuff
                end
              end
 
-    result.line lineno if lineno
-
     raise "identifier #{id.inspect} is not valid" unless result
 
     result
   end
 
+  def hack_encoding str, extra = nil
+    encodings = ENCODING_ORDER.dup
+    encodings.unshift(extra) unless extra.nil?
+
+    # terrible, horrible, no good, very bad, last ditch effort.
+    encodings.each do |enc|
+      begin
+        str.force_encoding enc
+        if str.valid_encoding? then
+          str.encode! Encoding::UTF_8
+          break
+        end
+      rescue ArgumentError # unknown encoding name
+        # do nothing
+      rescue Encoding::InvalidByteSequenceError
+        # do nothing
+      rescue Encoding::UndefinedConversionError
+        # do nothing
+      end
+    end
+
+    # no amount of pain is enough for you.
+    raise "Bad encoding. Need a magic encoding comment." unless
+      str.encoding.name == "UTF-8"
+  end
+
   ##
-  # Canonicalize conditionals. Eg:
-  #
-  #   not x ? a : b
+  # Returns a UTF-8 encoded string after processing BOMs and magic
+  # encoding comments.
   #
-  # becomes:
+  # Holy crap... ok. Here goes:
   #
-  #   x ? b : a
+  # Ruby's file handling and encoding support is insane. We need to be
+  # able to lex a file. The lexer file is explicitly UTF-8 to make
+  # things cleaner. This allows us to deal with extended chars in
+  # class and method names. In order to do this, we need to encode all
+  # input source files as UTF-8. First, we look for a UTF-8 BOM by
+  # looking at the first line while forcing its encoding to
+  # ASCII-8BIT. If we find a BOM, we strip it and set the expected
+  # encoding to UTF-8. Then, we search for a magic encoding comment.
+  # If found, it overrides the BOM. Finally, we force the encoding of
+  # the input string to whatever was found, and then encode that to
+  # UTF-8 for compatibility with the lexer.
+
+  def handle_encoding str
+    str = str.dup
+    has_enc = str.respond_to? :encoding # TODO: remove
+    encoding = nil
+
+    header = str.each_line.first(2)
+    header.map! { |s| s.force_encoding "ASCII-8BIT" } if has_enc
+
+    first = header.first || ""
+    encoding, str = +"utf-8", str.b[3..-1] if first =~ /\A\xEF\xBB\xBF/
 
-  attr_accessor :canonicalize_conditions
+    encoding = $1.strip if header.find { |s|
+      s[/^#.*?-\*-.*?coding:\s*([^ ;]+).*?-\*-/, 1] ||
+      s[/^#.*(?:en)?coding(?:\s*[:=])\s*([\w-]+)/, 1]
+    }
 
-  def initialize(options = {})
-    super()
+    if encoding then
+      if has_enc then
+        encoding.sub!(/utf-8-.+$/, "utf-8") # HACK for stupid emacs formats
+        hack_encoding str, encoding
+      else
+        warn "Skipping magic encoding comment"
+      end
+    else
+      # nothing specified... ugh. try to encode as utf-8
+      hack_encoding str if has_enc
+    end
 
-    v = self.class.name[/2\d/]
-    raise "Bad Class name #{self.class}" unless v
+    str
+  end
 
-    self.lexer = RubyLexer.new v && v.to_i
-    self.lexer.parser = self
-    self.in_kwarg = false
+  def invert_block_call val
+    ret, iter = val
+    type, call = ret
 
-    @env = RubyParserStuff::Environment.new
-    @comments = []
+    iter.insert 1, call
 
-    @canonicalize_conditions = true
+    ret = s(type).line ret.line
 
-    self.reset
+    [iter, ret]
+  end
+
+  def inverted? val
+    JUMP_TYPE[val[0].sexp_type]
   end
 
   def list_append list, item # TODO: nuke me *sigh*
@@ -350,12 +629,14 @@ module RubyParserStuff
 
     htype, ttype = head.sexp_type, tail.sexp_type
 
-    head = s(:dstr, '', head) if htype == :evstr
+    head = s(:dstr, "", head).line head.line if htype == :evstr
 
     case ttype
     when :str then
       if htype == :str
-        head.last << tail.last
+        a, b = head.last, tail.last
+        b = b.dup.force_encoding a.encoding unless Encoding.compatible?(a, b)
+        a << b
       elsif htype == :dstr and head.size == 2 then
         head.last << tail.last
       else
@@ -369,8 +650,8 @@ module RubyParserStuff
         head.line = lineno
       else
         tail.sexp_type = :array
-        tail[1] = s(:str, tail[1])
-        tail.delete_at 1 if tail[1] == s(:str, '')
+        tail[1] = s(:str, tail[1]).line tail.line
+        tail.delete_at 1 if tail[1] == s(:str, "")
 
         head.push(*tail.sexp_body)
       end
@@ -396,6 +677,13 @@ module RubyParserStuff
     return head
   end
 
+  def local_pop in_def
+    lexer.cond.pop # group = local_pop
+    lexer.cmdarg.pop
+    self.env.unextend
+    self.in_def = in_def
+  end
+
   def logical_op type, left, right
     left = value_expr left
 
@@ -408,12 +696,15 @@ module RubyParserStuff
         node = rhs
       end
 
-      node[2] = s(type, rhs, right)
+      node.pop
+      node << s(type, rhs, right).line(rhs.line)
 
       return left
     end
 
-    return s(type, left, right)
+    result = s(type, left, right)
+    result.line left.line if left.line
+    result
   end
 
   def new_aref val
@@ -422,6 +713,126 @@ module RubyParserStuff
     new_call val[0], :"[]", val[2]
   end
 
+  def new_arg val
+    arg, = val
+
+    case arg
+    when Symbol then
+      result = s(:args, arg).line line
+    when Sexp then
+      result = arg
+    when Array then
+      (arg, line), = val
+      result = s(:args, arg).line line
+    else
+      debug20 32
+      raise "Unknown f_arg type: #{val.inspect}"
+    end
+
+    result
+  end
+
+  def ary_to_pat ary
+    pat = ary.dup
+    pat.sexp_type = :array_TAIL
+
+    new_array_pattern nil, nil, pat, ary.line
+  end
+
+  def new_array_pattern const, pre_arg, arypat, loc
+    result = s(:array_pat, const).line loc
+    result << pre_arg if pre_arg
+
+    if arypat && arypat.sexp_type == :array_TAIL then
+      result.concat arypat.sexp_body
+    else
+      raise "NO?: %p" % [arypat]
+    end
+
+    result
+  end
+
+  def array_pat_concat lhs, rhs
+    case lhs.sexp_type
+    when :PATTERN then
+      lhs.sexp_type = :array_pat
+    end
+
+    if rhs then
+      case rhs.sexp_type
+      when :array_pat, :array_TAIL, :PATTERN then
+        lhs.concat rhs.sexp_body
+      else
+        lhs << rhs
+      end
+    end
+  end
+
+  def new_array_pattern_tail pre_args, has_rest, rest_arg, post_args
+    # TODO: remove has_rest once all tests pass !!!
+    rest_arg = if has_rest then
+                 :"*#{rest_arg}"
+               else
+                 nil
+               end
+
+    result = s(:array_TAIL).line 666
+
+    array_pat_concat result, pre_args
+
+    result << rest_arg if rest_arg
+
+    array_pat_concat result, post_args
+
+    result
+  end
+
+  def new_assign lhs, rhs
+    return nil unless lhs
+
+    rhs = value_expr rhs
+
+    case lhs.sexp_type
+    when :lasgn, :iasgn, :cdecl, :cvdecl, :gasgn, :cvasgn, :attrasgn, :safe_attrasgn then
+      lhs << rhs
+      lhs.line_max = rhs.line_max
+    when :const then
+      lhs.sexp_type = :cdecl
+      lhs << rhs
+    else
+      raise "unknown lhs #{lhs.inspect} w/ #{rhs.inspect}"
+    end
+
+    lhs
+  end
+
+  def new_attrasgn recv, meth, call_op = :"."
+    call_op = call_op.first if Array === call_op
+
+    meth = :"#{meth}="
+
+    result = case call_op.to_sym
+             when :"."
+               s(:attrasgn, recv, meth)
+             when :"&."
+               s(:safe_attrasgn, recv, meth)
+             else
+               raise "unknown call operator: `#{type.inspect}`"
+             end
+
+    result.line = recv.line
+    result
+  end
+
+  def new_begin val
+    (_, line), _, body, _ = val
+
+    result = body ? s(:begin, body) : s(:nil)
+    result.line line
+
+    result
+  end
+
   def new_body val
     body, resbody, elsebody, ensurebody = val
 
@@ -445,43 +856,29 @@ module RubyParserStuff
 
     if elsebody and not resbody then
       warning("else without rescue is useless")
-      result = s(:begin, result) if result
+      result = s(:begin, result).line result.line if result
       result = block_append(result, elsebody)
     end
 
-    result = s(:ensure, result, ensurebody).compact if ensurebody
+    if ensurebody
+      lineno = (result || ensurebody).line
+      result = s(:ensure, result, ensurebody).compact.line lineno
+    end
 
     result
   end
 
   def new_brace_body args, body, lineno
-    new_iter(nil, args, body).line(lineno)
+    new_iter(nil, args, body).line lineno
   end
 
-  def argl x
-    x = s(:arglist, x) if x and x.sexp_type == :array
-    x
-  end
-
-  def backref_assign_error ref
-    # TODO: need a test for this... obviously
-    case ref.sexp_type
-    when :nth_ref then
-      raise "write a test 2"
-      raise SyntaxError, "Can't set variable %p" % ref.last
-    when :back_ref then
-      raise "write a test 3"
-      raise SyntaxError, "Can't set back reference %p" % ref.last
-    else
-      raise "Unknown backref type: #{ref.inspect}"
-    end
-  end
+  def new_call recv, meth, args = nil, call_op = :"."
+    call_op = call_op.first if Array === call_op
 
-  def new_call recv, meth, args = nil, call_op = :'.'
     result = case call_op.to_sym
-             when :'.'
+             when :"."
                s(:call, recv, meth)
-             when :'&.'
+             when :"&."
                s(:safe_call, recv, meth)
              else
                raise "unknown call operator: `#{type.inspect}`"
@@ -490,40 +887,30 @@ module RubyParserStuff
     # TODO: need a test with f(&b) to produce block_pass
     # TODO: need a test with f(&b) { } to produce warning
 
-    if args
-      if [:arglist, :args, :array, :call_args].include? args.sexp_type
+    if args then
+      if ARG_TYPES[args.sexp_type] then
         result.concat args.sexp_body
       else
         result << args
       end
+      result.line_max = args.line_max
     end
 
-    line = result.grep(Sexp).map(&:line).compact.min
-    result.line = line if line
+    # line = result.grep(Sexp).map(&:line).compact.min
+    result.line = recv.line if recv
+    result.line ||= lexer.lineno
 
     result
   end
 
-  def new_attrasgn recv, meth, call_op
-    meth = :"#{meth}="
-
-    result = case call_op.to_sym
-             when :'.'
-               s(:attrasgn, recv, meth)
-             when :'&.'
-               s(:safe_attrasgn, recv, meth)
-             else
-               raise "unknown call operator: `#{type.inspect}`"
-             end
-
-    result.line = recv.line
-    result
+  def new_in pat, body, cases, line
+    s(:in, pat, body, cases).line line
   end
 
   def new_case expr, body, line
     result = s(:case, expr)
 
-    while body and body.node_type == :when
+    while body and [:when, :in].include? body.sexp_type
       result << body
       body = body.delete_at 3
     end
@@ -542,7 +929,9 @@ module RubyParserStuff
   end
 
   def new_class val
-    line, path, superclass, body = val[1], val[2], val[3], val[5]
+    (_, line, comment), path, superclass, _, body, (_, line_max) = val
+
+    path = path.first if path.instance_of? Array
 
     result = s(:class, path, superclass)
 
@@ -555,7 +944,8 @@ module RubyParserStuff
     end
 
     result.line = line
-    result.comments = self.comments.pop
+    result.line_max = line_max
+    result.comments = comment if comment
     result
   end
 
@@ -565,50 +955,147 @@ module RubyParserStuff
     result
   end
 
+  def new_const_op_asgn val
+    lhs, (asgn_op, _), rhs = val
+    asgn_op = asgn_op.to_sym
+
+    result = case asgn_op
+             when :"||" then
+               s(:op_asgn_or, lhs, rhs)
+             when :"&&" then
+               s(:op_asgn_and, lhs, rhs)
+             else
+               s(:op_asgn, lhs, asgn_op, rhs)
+             end
+
+    result.line = lhs.line
+    result
+  end
+
   def new_defn val
-    (_, line), (name, _), _, args, body, * = val
-    body ||= s(:nil)
+    if val.size == 4 then
+      ((_, line, comment), (name, _line, in_def)), args, body, (_, line_max) = val
+    else
+      (_, line, comment), (name, line), in_def, args, body, (_, line_max) = val
+    end
 
-    result = s(:defn, name.to_sym, args)
+    body ||= s(:nil).line line
 
-    if body then
-      if body.sexp_type == :block then
-        result.push(*body.sexp_body)
+    args.line line
+
+    result = s(:defn, name.to_sym, args).line line
+    result.line_max = line_max
+
+    if body.sexp_type == :block then
+      result.push(*body.sexp_body)
+    else
+      result.push body
+    end
+
+    result.comments = comment if comment
+
+    [result, in_def]
+  end
+
+  def new_endless_defn val
+    # not available in 2.x so we don't need to check size
+    ((_, line, comment), (name, _, in_def)), args, _, body, _, resbody = val
+
+    result =
+      if resbody then
+        s(:defn, name, args,
+          new_rescue(body,
+                     new_resbody(s(:array).line(line),
+                                 resbody))).line line
       else
-        result.push body
+        s(:defn, name, args, body).line line
       end
-    end
 
-    args.line line
-    result.line = line
-    result.comments = self.comments.pop
+    local_pop in_def
+    endless_method_name result
+
+    result.comments = comment if comment
 
     result
   end
 
-  def new_defs val
-    recv, (name, _line), args, body = val[1], val[4], val[6], val[7]
-    body ||= s(:nil)
-
-    result = s(:defs, recv, name.to_sym, args)
+  def new_endless_defs val
+    # not available in 2.x so we don't need to check size
+    ((_, line, comment), recv, _, _, (name, line, in_def)), \
+      args, _, body, _, resbody = val
 
-    if body then
-      if body.sexp_type == :block then
-        result.push(*body.sexp_body)
+    result =
+      if resbody then
+        s(:defs, recv, name, args,
+          new_rescue(body,
+                     new_resbody(s(:array).line(line),
+                                 resbody))).line line
       else
-        result.push body
+        s(:defs, recv, name, args, body).line(line)
       end
-    end
 
-    result.line = recv.line
-    result.comments = self.comments.pop
+    self.in_single -= 1
+    local_pop in_def
+    endless_method_name result
+
+    result.comments = comment if comment
+
     result
   end
 
+  def new_defs val
+    if val.size == 4 then
+      ((_, line, comment), recv, _, _, (name, line, in_def)), \
+        args, body, (_, line_max) = val
+    else
+      (_, line, comment), recv, (name, _), in_def, \
+        args, body, (_, line_max) = val
+    end
+
+    body ||= s(:nil).line line
+
+    args.line line
+
+    result = s(:defs, recv, name.to_sym, args).line line
+    result.line_max = line_max
+
+    # TODO: remove_begin
+    # TODO: reduce_nodes
+
+    if body.sexp_type == :block then
+      result.push(*body.sexp_body)
+    else
+      result.push body
+    end
+
+    result.comments = comment if comment
+
+    [result, in_def]
+  end
+
   def new_do_body args, body, lineno
     new_iter(nil, args, body).line(lineno)
   end
 
+  def new_find_pattern const, pat
+    pat.sexp_type = :find_pat
+    pat.insert 1, const
+  end
+
+  def new_find_pattern_tail lhs, mid, rhs
+    lhs_id, line = lhs
+    rhs_id, _line = rhs
+
+    # TODO: fpinfo->pre_rest_arg = pre_rest_arg ? assignable(p, pre_rest_arg, 0, loc) : NODE_SPECIAL_NO_NAME_REST;
+
+    lhs_id = "*#{lhs_id}".to_sym
+    rhs_id = "*#{rhs_id}".to_sym
+
+    raise "BAD?" unless mid.sexp_type == :array_TAIL
+
+    s(:find_pat_TAIL, lhs_id, *mid.sexp_body, rhs_id).line line
+  end
+
   def new_for expr, var, body
     result = s(:for, expr, var).line(var.line)
     result << body if body
@@ -616,7 +1103,49 @@ module RubyParserStuff
   end
 
   def new_hash val
-    s(:hash, *val[2].values).line(val[1])
+    _, line, assocs = val
+
+    s(:hash).line(line).concat assocs.sexp_body
+  end
+
+  def new_hash_pattern const, hash_pat, loc
+    _, pat, kw_args, kw_rest_arg = hash_pat
+
+    line = (const||hash_pat).line
+
+    result = s(:hash_pat, const).line line
+    result.concat pat.sexp_body if pat
+    result << kw_args     if kw_args
+    result << kw_rest_arg if kw_rest_arg
+    result
+  end
+
+  def new_hash_pattern_tail kw_args, kw_rest_arg, line # TODO: remove line arg
+    # kw_rest_arg = assignable(kw_rest_arg, nil).line line if kw_rest_arg
+
+    result = s(:hash_pat).line line
+    result << kw_args
+
+    if kw_rest_arg then
+      name = kw_rest_arg.value
+      # TODO: I _hate_ this:
+      assignable [name, kw_rest_arg.line] if name != :**
+      result << kw_rest_arg
+    end
+
+    result
+  end
+
+  def push_pktbl
+  end
+
+  def pop_pktbl
+  end
+
+  def push_pvtbl
+  end
+
+  def pop_pvtbl
   end
 
   def new_if c, t, f
@@ -637,33 +1166,68 @@ module RubyParserStuff
     result << args
     result << body if body
 
-    args.sexp_type = :args unless args == 0
+    result.line call.line if call
 
-    result
-  end
+    unless args == 0 then
+      args.line call.line if call
+      args.sexp_type = :args
+    end
 
-  def new_masgn_arg rhs, wrap = false
-    rhs = value_expr(rhs)
-    rhs = s(:to_ary, rhs) if wrap # HACK: could be array if lhs isn't right
-    rhs
+    result
   end
 
   def new_masgn lhs, rhs, wrap = false
     _, ary = lhs
 
+    line = rhs.line
     rhs = value_expr(rhs)
     rhs = ary ? s(:to_ary, rhs) : s(:array, rhs) if wrap
+    rhs.line line if wrap
 
     lhs.delete_at 1 if ary.nil?
     lhs << rhs
 
-    lhs
+    lhs
+  end
+
+  def new_masgn_arg rhs, wrap = false
+    rhs = value_expr(rhs)
+    # HACK: could be array if lhs isn't right
+    rhs = s(:to_ary, rhs).line rhs.line if wrap
+    rhs
+  end
+
+  def new_match lhs, rhs
+    if lhs then
+      case lhs.sexp_type
+      when :dregx, :dregx_once then
+        # TODO: no test coverage
+        return s(:match2, lhs, rhs).line(lhs.line)
+      when :lit then
+        return s(:match2, lhs, rhs).line(lhs.line) if Regexp === lhs.last
+      end
+    end
+
+    if rhs then
+      case rhs.sexp_type
+      when :dregx, :dregx_once then
+        # TODO: no test coverage
+        return s(:match3, rhs, lhs).line(lhs.line)
+      when :lit then
+        return s(:match3, rhs, lhs).line(lhs.line) if Regexp === rhs.last
+      end
+    end
+
+    new_call(lhs, :"=~", argl(rhs)).line lhs.line
   end
 
   def new_module val
-    line, path, body = val[1], val[2], val[4]
+    (_, line_min, comment), path, _, body, (_, line_max) = val
 
-    result = s(:module, path)
+    path = path.first if path.instance_of? Array
+
+    result = s(:module, path).line line_min
+    result.line_max = line_max
 
     if body then # REFACTOR?
       if body.sexp_type == :block then
@@ -673,55 +1237,49 @@ module RubyParserStuff
       end
     end
 
-    result.line = line
-    result.comments = self.comments.pop
+    result.comments = comment if comment
     result
   end
 
   def new_op_asgn val
-    lhs, asgn_op, arg = val[0], val[1].to_sym, val[2]
-    name = lhs.value
-    arg = remove_begin(arg)
-    result = case asgn_op # REFACTOR
+    lhs, (op, _line), rhs = val
+    op = op.to_sym
+
+    name = gettable(lhs.last).line lhs.line
+    arg = remove_begin rhs
+    result = case op # REFACTOR
              when :"||" then
                lhs << arg
-               s(:op_asgn_or, self.gettable(name), lhs)
+               s(:op_asgn_or, name, lhs).line lhs.line
              when :"&&" then
                lhs << arg
-               s(:op_asgn_and, self.gettable(name), lhs)
+               s(:op_asgn_and, name, lhs).line lhs.line
              else
-               # TODO: why [2] ?
-               lhs[2] = new_call(self.gettable(name), asgn_op, argl(arg))
+               lhs << new_call(name, op, argl(arg))
                lhs
              end
-    result.line = lhs.line
+
     result
   end
 
-  def new_const_op_asgn val
-    lhs, asgn_op, rhs = val[0], val[1].to_sym, val[2]
+  def new_op_asgn1 val
+    lhs, _, args, _, (op, _), rhs = val
 
-    result = case asgn_op
-             when :"||" then
-               s(:op_asgn_or, lhs, rhs)
-             when :"&&" then
-               s(:op_asgn_and, lhs, rhs)
-             else
-               s(:op_asgn, lhs, asgn_op, rhs)
-             end
+    args.sexp_type = :arglist if args
 
-    result.line = lhs.line
+    result = s(:op_asgn1, lhs, args, op.to_sym, rhs)
+    result.line lhs.line
     result
   end
 
   def new_op_asgn2 val
-    recv, call_op, meth, op, arg = val
+    recv, (call_op, _), (meth, _), (op, _), arg = val
     meth = :"#{meth}="
 
     result = case call_op.to_sym
-             when :'.'
+             when :"."
                s(:op_asgn2, recv, meth, op.to_sym, arg)
-             when :'&.'
+             when :"&."
                s(:safe_op_asgn2, recv, meth, op.to_sym, arg)
              else
                raise "unknown call operator: `#{type.inspect}`"
@@ -731,21 +1289,42 @@ module RubyParserStuff
     result
   end
 
+  def new_qsym_list
+    s(:array).line lexer.lineno
+  end
+
+  def new_qsym_list_entry val
+    _, (str, line), _ = val
+    s(:lit, str.to_sym).line line
+  end
+
+  def new_qword_list
+    s(:array).line lexer.lineno
+  end
+
+  def new_qword_list_entry val
+    _, (str, line), _ = val
+    str.force_encoding("ASCII-8BIT") unless str.valid_encoding?
+    s(:str, str).line line
+  end
+
   def new_regexp val
-    node = val[1] || s(:str, '')
-    options = val[2]
+    (_, line), node, (options, line_max) = val
+
+    node ||= s(:str, "").line line
+    node.line_max = line_max
 
     o, k = 0, nil
     options.split(//).uniq.each do |c| # FIX: this has a better home
       v = {
-        'x' => Regexp::EXTENDED,
-        'i' => Regexp::IGNORECASE,
-        'm' => Regexp::MULTILINE,
-        'o' => Regexp::ONCE,
-        'n' => Regexp::ENC_NONE,
-        'e' => Regexp::ENC_EUC,
-        's' => Regexp::ENC_SJIS,
-        'u' => Regexp::ENC_UTF8,
+        "x" => Regexp::EXTENDED,
+        "i" => Regexp::IGNORECASE,
+        "m" => Regexp::MULTILINE,
+        "o" => Regexp::ONCE,
+        "n" => Regexp::ENC_NONE,
+        "e" => Regexp::ENC_EUC,
+        "s" => Regexp::ENC_SJIS,
+        "u" => Regexp::ENC_UTF8,
       }[c]
       raise "unknown regexp option: #{c}" unless v
       o += v
@@ -760,12 +1339,12 @@ module RubyParserStuff
                   begin
                     Regexp.new(node[1], o)
                   rescue RegexpError => e
-                    warn "WA\RNING: #{e.message} for #{node[1].inspect} #{options.inspect}"
+                    warn "WARNING: #{e.message} for #{node[1].inspect} #{options.inspect}"
                     begin
-                      warn "WA\RNING: trying to recover with ENC_UTF8"
+                      warn "WARNING: trying to recover with ENC_UTF8"
                       Regexp.new(node[1], Regexp::ENC_UTF8)
                     rescue RegexpError => e
-                      warn "WA\RNING: trying to recover with ENC_NONE"
+                      warn "WARNING: trying to recover with ENC_NONE"
                       Regexp.new(node[1], Regexp::ENC_NONE)
                     end
                   end
@@ -778,7 +1357,7 @@ module RubyParserStuff
       end
       node << o if o and o != 0
     else
-      node = s(:dregx, '', node);
+      node = s(:dregx, "", node).line line
       node.sexp_type = :dregx_once if options =~ /o/
       node << o if o and o != 0
     end
@@ -786,21 +1365,22 @@ module RubyParserStuff
     node
   end
 
-  def new_rescue body, resbody
-    s(:rescue, body, resbody)
-  end
-
   def new_resbody cond, body
     if body && body.sexp_type == :block then
       body.shift # remove block and splat it in directly
     else
       body = [body]
     end
+
     s(:resbody, cond, *body).line cond.line
   end
 
+  def new_rescue body, resbody
+    s(:rescue, body, resbody).line body.line
+  end
+
   def new_sclass val
-    recv, in_def, in_single, body = val[3], val[4], val[6], val[7]
+    (_, line), _, recv, in_def, _, in_single, body, _ = val
 
     result = s(:sclass, recv)
 
@@ -812,97 +1392,62 @@ module RubyParserStuff
       end
     end
 
-    result.line = val[2]
+    result.line = line
     self.in_def = in_def
     self.in_single = in_single
     result
   end
 
   def new_string val
-    str = val[0]
-    str.force_encoding("UTF-8")
-    str.force_encoding("ASCII-8BIT") unless str.valid_encoding?
-    result = s(:str, str)
-    self.lexer.fixup_lineno str.count("\n")
-    result
-  end
+    (str, line), = val
 
-  def new_qword_list_entry val
-    str = val[1]
+    str.force_encoding("UTF-8")
+    # TODO: remove:
     str.force_encoding("ASCII-8BIT") unless str.valid_encoding?
-    result = s(:str, str)
-    self.lexer.fixup_lineno
-    result
-  end
-
-  def new_qword_list
-    result = s(:array)
-    self.lexer.fixup_lineno
-    result
-  end
-
-  def new_word_list
-    result = s(:array)
-    self.lexer.fixup_lineno
-    result
-  end
-
-  def new_word_list_entry val
-    result = val[1].sexp_type == :evstr ? s(:dstr, "", val[1]) : val[1]
-    self.lexer.fixup_lineno
-    result
+    s(:str, str).line line
   end
 
-  def new_qsym_list
-    result = s(:array)
-    self.lexer.fixup_lineno
-    result
+  def new_super args
+    if args && args.sexp_type == :block_pass then
+      s(:super, args).line args.line
+    else
+      args ||= s(:arglist).line lexer.lineno
+      s(:super, *args.sexp_body).line args.line
+    end
   end
 
-  def new_qsym_list_entry val
-    result = s(:lit, val[1].to_sym)
-    self.lexer.fixup_lineno
-    result
+  def new_symbol val
+    name = val.last
+    s(:lit, name.to_sym).line lexer.lineno
   end
 
   def new_symbol_list
-    result = s(:array)
-    self.lexer.fixup_lineno
-    result
+    # TODO: hunt down and try to remove ALL lexer.lineno usage!
+    s(:array).line lexer.lineno
   end
 
   def new_symbol_list_entry val
-    _list, sym, _nil = val # TODO: use _list
-    result = val[1]
+    _, sym, _ = val
 
-    result ||= s(:str, "")
+    sym ||= s(:str, "").line lexer.lineno
 
     case sym.sexp_type
     when :dstr then
       sym.sexp_type = :dsym
     when :str then
-      sym = s(:lit, sym.last.to_sym)
+      sym = s(:lit, sym.last.to_sym).line sym.line
     else
-      sym = s(:dsym, "", sym || s(:str, ""))
+      sym = s(:dsym, "", sym).line sym.line
     end
-    self.lexer.fixup_lineno
-    sym
-  end
 
-  def new_super args
-    if args && args.node_type == :block_pass then
-      s(:super, args)
-    else
-      args ||= s(:arglist)
-      s(:super, *args.sexp_body)
-    end
+    sym
   end
 
   def new_undef n, m = nil
     if m then
-      block_append(n, s(:undef, m))
+      block_append(n, s(:undef, m).line(m.line))
     else
-      s(:undef, n)
+      s(:undef, n).line n.line
     end
   end
 
@@ -935,135 +1480,83 @@ module RubyParserStuff
     new_until_or_while :while, block, expr, pre
   end
 
-  def new_xstring str
-    if str then
-      case str.sexp_type
+  def new_word_list
+    s(:array).line lexer.lineno
+  end
+
+  def new_word_list_entry val
+    _, word, _ = val
+    word.sexp_type == :evstr ? s(:dstr, "", word).line(word.line) : word
+  end
+
+  def new_xstring val
+    _, node = val
+
+    node ||= s(:str, "").line lexer.lineno
+
+    if node then
+      case node.sexp_type
       when :str
-        str.sexp_type = :xstr
+        node.sexp_type = :xstr
       when :dstr
-        str.sexp_type = :dxstr
+        node.sexp_type = :dxstr
       else
-        str = s(:dxstr, '', str)
+        node = s(:dxstr, "", node).line node.line
       end
-      str
-    else
-      s(:xstr, '')
     end
+
+    node
   end
 
   def new_yield args = nil
     # TODO: raise args.inspect unless [:arglist].include? args.first # HACK
-    raise "write a test 4" if args && args.node_type == :block_pass
+    raise "write a test 4" if args && args.sexp_type == :block_pass
     raise SyntaxError, "Block argument should not be given." if
-      args && args.node_type == :block_pass
+      args && args.sexp_type == :block_pass
 
-    args ||= s(:arglist)
+    args ||= s(:arglist).line lexer.lineno
 
     args.sexp_type = :arglist if [:call_args, :array].include? args.sexp_type
-    args = s(:arglist, args) unless args.sexp_type == :arglist
-
-    return s(:yield, *args.sexp_body)
-  end
-
-  def next_token
-    token = self.lexer.next_token
+    args = s(:arglist, args).line args.line unless args.sexp_type == :arglist
 
-    if token and token.first != RubyLexer::EOF then
-      return token
-    else
-      return [false, '$end']
-    end
+    s(:yield, *args.sexp_body).line args.line
   end
 
-  def new_assign lhs, rhs
-    return nil unless lhs
-
-    rhs = value_expr rhs
-
-    case lhs.sexp_type
-    when :lasgn, :iasgn, :cdecl, :cvdecl, :gasgn, :cvasgn, :attrasgn, :safe_attrasgn then
-      lhs << rhs
-    when :const then
-      lhs.sexp_type = :cdecl
-      lhs << rhs
+  def prev_value_to_lineno v
+    s, n = v
+    if String === s then
+      n
     else
-      raise "unknown lhs #{lhs.inspect} w/ #{rhs.inspect}"
+      lexer.lineno
     end
-
-    lhs
   end
 
-  ##
-  # Returns a UTF-8 encoded string after processing BOMs and magic
-  # encoding comments.
-  #
-  # Holy crap... ok. Here goes:
-  #
-  # Ruby's file handling and encoding support is insane. We need to be
-  # able to lex a file. The lexer file is explicitly UTF-8 to make
-  # things cleaner. This allows us to deal with extended chars in
-  # class and method names. In order to do this, we need to encode all
-  # input source files as UTF-8. First, we look for a UTF-8 BOM by
-  # looking at the first line while forcing its encoding to
-  # ASCII-8BIT. If we find a BOM, we strip it and set the expected
-  # encoding to UTF-8. Then, we search for a magic encoding comment.
-  # If found, it overrides the BOM. Finally, we force the encoding of
-  # the input string to whatever was found, and then encode that to
-  # UTF-8 for compatibility with the lexer.
-
-  def handle_encoding str
-    str = str.dup
-    has_enc = str.respond_to? :encoding
-    encoding = nil
+  KEEP_COMMENT_TOKENS = [:kCLASS, :kMODULE, :kDEF, :tNL]
 
-    header = str.each_line.first(2)
-    header.map! { |s| s.force_encoding "ASCII-8BIT" } if has_enc
+  def next_token
+    token = self.lexer.next_token
 
-    first = header.first || ""
-    encoding, str = "utf-8", str[3..-1] if first =~ /\A\xEF\xBB\xBF/
+    if token and token.first != RubyLexer::EOF then
+      self.last_token_type = token
 
-    encoding = $1.strip if header.find { |s|
-      s[/^#.*?-\*-.*?coding:\s*([^ ;]+).*?-\*-/, 1] ||
-      s[/^#.*(?:en)?coding(?:\s*[:=])\s*([\w-]+)/, 1]
-    }
+      self.lexer.comment = nil unless KEEP_COMMENT_TOKENS.include? token.first
 
-    if encoding then
-      if has_enc then
-        encoding.sub!(/utf-8-.+$/, 'utf-8') # HACK for stupid emacs formats
-        hack_encoding str, encoding
-      else
-        warn "Skipping magic encoding comment"
-      end
+      return token
+    elsif !token
+      return self.lexer.next_token
     else
-      # nothing specified... ugh. try to encode as utf-8
-      hack_encoding str if has_enc
+      return [false, false]
     end
-
-    str
   end
 
-  def hack_encoding str, extra = nil
-    encodings = ENCODING_ORDER.dup
-    encodings.unshift(extra) unless extra.nil?
-
-    # terrible, horrible, no good, very bad, last ditch effort.
-    encodings.each do |enc|
-      begin
-        str.force_encoding enc
-        if str.valid_encoding? then
-          str.encode! Encoding::UTF_8
-          break
-        end
-      rescue Encoding::InvalidByteSequenceError
-        # do nothing
-      rescue Encoding::UndefinedConversionError
-        # do nothing
-      end
-    end
-
-    # no amount of pain is enough for you.
-    raise "Bad encoding. Need a magic encoding comment." unless
-      str.encoding.name == "UTF-8"
+  def on_error(et, ev, values)
+    ev = ev.first if ev.instance_of?(Array) && ev.size == 2 && ev.last.is_a?(Integer)
+    super
+  rescue Racc::ParseError => e
+    # I don't like how the exception obscures the error message
+    e.message.replace "%s:%p :: %s" % [self.file, lexer.lineno, e.message.strip]
+    warn e.message if $DEBUG
+    raise
   end
 
   ##
@@ -1071,56 +1564,43 @@ module RubyParserStuff
   # Timeout::Error if it runs for more than +time+ seconds.
 
   def process(str, file = "(string)", time = 10)
+    str.freeze
+
     Timeout.timeout time do
       raise "bad val: #{str.inspect}" unless String === str
 
-      str = handle_encoding str
+      self.lexer.string = handle_encoding str
 
       self.file = file.dup
 
-      @yydebug = ENV.has_key? 'DEBUG'
-
-      # HACK -- need to get tests passing more than have graceful code
-      self.lexer.ss = RPStringScanner.new str
+      @yydebug = ENV.has_key? "DEBUG"
 
       do_parse
     end
   end
 
-  alias :parse :process
+  alias parse process
 
   def remove_begin node
-    oldnode = node
-    if node and node.sexp_type == :begin and node.size == 2 then
-      node = node.last
-      node.line = oldnode.line
-    end
+    line = node.line
+
+    node = node.last while node and node.sexp_type == :begin and node.size == 2
+
+    node = s(:nil) if node == s(:begin)
+
+    node.line ||= line
+
     node
   end
 
+  alias value_expr remove_begin # TODO: for now..? could check the tree, but meh?
+
   def reset
     lexer.reset
     self.in_def = false
     self.in_single = 0
     self.env.reset
-    self.comments.clear
-  end
-
-  def block_dup_check call_or_args, block
-    syntax_error "Both block arg and actual block given." if
-      block and call_or_args.block_pass?
-  end
-
-  def inverted? val
-    [:return, :next, :break, :yield].include? val[0].sexp_type
-  end
-
-  def invert_block_call val
-    (type, call), iter = val
-
-    iter.insert 1, call
-
-    [iter, s(type)]
+    self.last_token_type = nil
   end
 
   def ret_args node
@@ -1135,7 +1615,7 @@ module RubyParserStuff
 
       # HACK matz wraps ONE of the FOUR splats in a newline to
       # distinguish. I use paren for now. ugh
-      node = s(:svalue, node) if node.sexp_type == :splat and not node.paren
+      node = s(:svalue, node).line node.line if node.sexp_type == :splat and not node.paren
       node.sexp_type = :svalue if node.sexp_type == :arglist && node[1].sexp_type == :splat
     end
 
@@ -1144,18 +1624,25 @@ module RubyParserStuff
 
   def s(*args)
     result = Sexp.new(*args)
-    result.line ||= lexer.lineno if lexer.ss          # otherwise...
+    # result.line ||= lexer.lineno if lexer.ss unless ENV["CHECK_LINE_NUMS"] # otherwise...
     result.file = self.file
     result
   end
 
-  def value_expr oldnode # HACK: much more to do
-    node = remove_begin oldnode
-    node.line = oldnode.line if oldnode
-    node[2] = value_expr node[2] if node and node.sexp_type == :if
-    node
+  def debug n
+    if ENV["PRY"] then
+      require "pry"; binding.pry
+    end
+
+    raise RubyParser::SyntaxError, "debug #{n}"
+  end
+
+  def syntax_error msg
+    raise RubyParser::SyntaxError, msg
   end
 
+  alias yyerror syntax_error
+
   def void_stmts node
     return nil unless node
     return node unless node.sexp_type == :block
@@ -1173,15 +1660,40 @@ module RubyParserStuff
     # do nothing for now
   end
 
-  alias yyerror syntax_error
+  def whitespace_width line, remove_width = nil
+    col = 0
+    idx = 0
 
-  def on_error(et, ev, values)
-    super
-  rescue Racc::ParseError => e
-    # I don't like how the exception obscures the error message
-    e.message.replace "%s:%p :: %s" % [self.file, lexer.lineno, e.message.strip]
-    warn e.message if $DEBUG
-    raise
+    line.chars.each do |c|
+      break if remove_width && col >= remove_width
+      case c
+      when " " then
+        col += 1
+      when "\t" then
+        n = TAB_WIDTH * (col / TAB_WIDTH + 1)
+        break if remove_width && n > remove_width
+        col = n
+      else
+        break
+      end
+      idx += 1
+    end
+
+    if remove_width then
+      line[idx..-1]
+    elsif line[idx] == "\n"
+      nil
+    else
+      col
+    end
+  end
+
+  alias remove_whitespace_width whitespace_width
+
+  def wrap type, node
+    value, line = node
+    value = value.to_sym if value.respond_to? :to_sym
+    s(type, value).line line
   end
 
   class Keyword
@@ -1214,8 +1726,10 @@ module RubyParserStuff
     # :expr_fitem   = symbol literal as FNAME.
     # :expr_value   = :expr_beg -- work to remove. Need multi-state support.
 
+    expr_woot = EXPR_FNAME|EXPR_FITEM
+
     wordlist = [
-                ["alias",    [:kALIAS,    :kALIAS      ], EXPR_FNAME|EXPR_FITEM],
+                ["alias",    [:kALIAS,    :kALIAS      ], expr_woot  ],
                 ["and",      [:kAND,      :kAND        ], EXPR_BEG   ],
                 ["begin",    [:kBEGIN,    :kBEGIN      ], EXPR_BEG   ],
                 ["break",    [:kBREAK,    :kBREAK      ], EXPR_MID   ],
@@ -1245,7 +1759,7 @@ module RubyParserStuff
                 ["super",    [:kSUPER,    :kSUPER      ], EXPR_ARG   ],
                 ["then",     [:kTHEN,     :kTHEN       ], EXPR_BEG   ],
                 ["true",     [:kTRUE,     :kTRUE       ], EXPR_END   ],
-                ["undef",    [:kUNDEF,    :kUNDEF      ], EXPR_FNAME|EXPR_FITEM],
+                ["undef",    [:kUNDEF,    :kUNDEF      ], expr_woot  ],
                 ["unless",   [:kUNLESS,   :kUNLESS_MOD ], EXPR_BEG   ],
                 ["until",    [:kUNTIL,    :kUNTIL_MOD  ], EXPR_BEG   ],
                 ["when",     [:kWHEN,     :kWHEN       ], EXPR_BEG   ],
@@ -1325,11 +1839,6 @@ module RubyParserStuff
       @debug = debug
     end
 
-    def reset
-      @stack = [false]
-      log :reset if debug
-    end
-
     def inspect
       "StackState(#{@name}, #{@stack.inspect})"
     end
@@ -1366,16 +1875,21 @@ module RubyParserStuff
       log :push if debug
     end
 
-    def store base = false
-      result = @stack.dup
-      @stack.replace [base]
-      log :store if debug
-      result
+    def reset
+      @stack = [false]
+      log :reset if debug
     end
 
     def restore oldstate
       @stack.replace oldstate
       log :restore if debug
     end
+
+    def store base = false
+      result = @stack.dup
+      @stack.replace [base]
+      log :store if debug
+      result
+    end
   end
 end