ruby_parser 3.13.1 → 3.15.1

data/lib/ruby_parser_extras.rb

@@ -1,4 +1,5 @@
 # encoding: ASCII-8BIT
+# TODO: remove
 
 require "sexp"
 require "ruby_lexer"
@@ -6,13 +7,50 @@ 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 maximum 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 = "3.15.1"
 
   attr_accessor :lexer, :in_def, :in_single, :file
   attr_accessor :in_kwarg
   attr_reader :env, :comments
 
+  ##
+  # 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 = []
 
   %w[
@@ -31,6 +69,28 @@ 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
+
   has_enc = "".respond_to? :encoding
 
   # This is in sorted order of occurrence according to
@@ -48,85 +108,52 @@ 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
-  end
+  TAB_WIDTH = 8
 
-  def arg_concat node1, node2 # TODO: nuke
-    raise "huh" unless node2
-    node1 << s(:splat, node2).compact
-    node1
-  end
+  def initialize(options = {})
+    super()
 
-  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
+    v = self.class.name[/2\d/]
+    raise "Bad Class name #{self.class}" unless v
 
-  def block_var *args
-    result = self.args args
-    result.sexp_type = :masgn
-    result
-  end
+    self.lexer = RubyLexer.new v && v.to_i
+    self.lexer.parser = self
+    self.in_kwarg = false
 
-  def array_to_hash array
-    case array.sexp_type
-    when :kwsplat then
-      array
-    else
-      s(:hash, *array.sexp_body)
-    end
+    @env = RubyParserStuff::Environment.new
+    @comments = []
+
+    @canonicalize_conditions = true
+
+    self.reset
   end
 
-  def call_args args
-    result = s(:call_args)
+  def arg_concat node1, node2 # TODO: nuke
+    raise "huh" unless node2
 
-    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}"
-      end
-    end
+    splat = s(:splat, node2)
+    splat.line node2.line
 
-    result
+    node1 << splat
+  end
+
+  def argl x
+    x = s(:arglist, x) if x and x.sexp_type == :array
+    x
   end
 
   def args args
     result = s(:args)
 
+    ss = args.grep Sexp
+    if ss.empty? then
+      result.line lexer.lineno
+    else
+      result.line ss.first.line
+    end
+
     args.each do |arg|
       case arg
       when Sexp then
@@ -162,13 +189,28 @@ module RubyParserStuff
     result
   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
+
+    raise "WTF" if Sexp === id
     id = id.to_sym if Sexp === id
 
     raise "write a test 1" if id.to_s =~ /^(?:self|nil|true|false|__LINE__|__FILE__)$/
@@ -197,11 +239,33 @@ module RubyParserStuff
 
     self.env[id] ||= :lvar if result.sexp_type == :lasgn
 
+    line = case lhs
+           when Sexp then
+             lhs.line
+           else
+             value && value.line || lexer.lineno
+           end
+
     result << value if value
+    result.line = line
 
     return result
   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 block_append(head, tail)
     return head if tail.nil?
     return tail if head.nil?
@@ -215,6 +279,69 @@ module RubyParserStuff
     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|
+      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}"
+      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
@@ -222,56 +349,113 @@ 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[/^[ \t]*/] }
+    }.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)
@@ -303,33 +487,94 @@ module RubyParserStuff
     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.
 
-  attr_accessor :canonicalize_conditions
+  def handle_encoding str
+    str = str.dup
+    has_enc = str.respond_to? :encoding
+    encoding = nil
 
-  def initialize(options = {})
-    super()
+    header = str.each_line.first(2)
+    header.map! { |s| s.force_encoding "ASCII-8BIT" } if has_enc
 
-    v = self.class.name[/2\d/]
-    raise "Bad Class name #{self.class}" unless v
+    first = header.first || ""
+    encoding, str = "utf-8", str.b[3..-1] if first =~ /\A\xEF\xBB\xBF/
 
-    self.lexer = RubyLexer.new v && v.to_i
-    self.lexer.parser = self
-    self.in_kwarg = false
+    encoding = $1.strip if header.find { |s|
+      s[/^#.*?-\*-.*?coding:\s*([^ ;]+).*?-\*-/, 1] ||
+      s[/^#.*(?:en)?coding(?:\s*[:=])\s*([\w-]+)/, 1]
+    }
 
-    @env = RubyParserStuff::Environment.new
-    @comments = []
+    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
 
-    @canonicalize_conditions = true
+    str
+  end
 
-    self.reset
+  def invert_block_call val
+    ret, iter = val
+    type, call = ret
+
+    iter.insert 1, call
+
+    ret = s(type).line ret.line
+
+    [iter, ret]
+  end
+
+  def inverted? val
+    JUMP_TYPE[val[0].sexp_type]
   end
 
   def list_append list, item # TODO: nuke me *sigh*
@@ -350,12 +595,14 @@ 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
@@ -369,15 +616,15 @@ 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
     when :evstr then
       if htype == :str then
         f, l = head.file, head.line
-        head = s(:dstr, *head.sexp_body)
+        head = s(:dstr, *head.sexp_body).line head.line
         head.file = f
         head.line = l
       end
@@ -408,12 +655,15 @@ 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 +672,49 @@ module RubyParserStuff
     new_call val[0], :"[]", val[2]
   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
+    else
+      raise "unknown lhs #{lhs.inspect} w/ #{rhs.inspect}"
+    end
+
+    lhs
+  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
+  end
+
+  def new_begin val
+    _, lineno, body, _ = val
+
+    result = body ? s(:begin, body) : s(:nil)
+    result.line lineno
+
+    result
+  end
+
   def new_body val
     body, resbody, elsebody, ensurebody = val
 
@@ -445,43 +738,27 @@ 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 = :'.'
+  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}`"
@@ -491,32 +768,17 @@ module RubyParserStuff
     # TODO: need a test with f(&b) { } to produce warning
 
     if args
-      if [:arglist, :args, :array, :call_args].include? args.sexp_type
+      if ARG_TYPES[args.sexp_type] then
         result.concat args.sexp_body
       else
         result << args
       end
     end
 
-    line = result.grep(Sexp).map(&:line).compact.min
-    result.line = line if line
-
-    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
+    # line = result.grep(Sexp).map(&:line).compact.min
+    result.line = recv.line if recv
+    result.line ||= lexer.lineno
 
-    result.line = recv.line
     result
   end
 
@@ -565,11 +827,29 @@ module RubyParserStuff
     result
   end
 
+  def new_const_op_asgn val
+    lhs, asgn_op, rhs = val[0], val[1].to_sym, val[2]
+
+    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)
+    (_, line), name, _, args, body, nil_body_line, * = val
+    body ||= s(:nil).line nil_body_line
 
-    result = s(:defn, name.to_sym, args)
+    args.line line
+
+    result = s(:defn, name.to_sym, args).line line
 
     if body then
       if body.sexp_type == :block then
@@ -579,19 +859,23 @@ module RubyParserStuff
       end
     end
 
-    args.line line
-    result.line = line
     result.comments = self.comments.pop
 
     result
   end
 
   def new_defs val
-    recv, (name, _line), args, body = val[1], val[4], val[6], val[7]
-    body ||= s(:nil)
+    _, recv, _, _, name, (_in_def, line), args, body, _ = val
+
+    body ||= s(:nil).line line
+
+    args.line line
 
     result = s(:defs, recv, name.to_sym, args)
 
+    # TODO: remove_begin
+    # TODO: reduce_nodes
+
     if body then
       if body.sexp_type == :block then
         result.push(*body.sexp_body)
@@ -616,7 +900,9 @@ module RubyParserStuff
   end
 
   def new_hash val
-    s(:hash, *val[2].values).line(val[1])
+    _, line, assocs = val
+
+    s(:hash).line(line).concat assocs.values
   end
 
   def new_if c, t, f
@@ -637,27 +923,59 @@ 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
@@ -680,37 +998,30 @@ module RubyParserStuff
 
   def new_op_asgn val
     lhs, asgn_op, arg = val[0], val[1].to_sym, val[2]
-    name = lhs.value
+    name = gettable(lhs.value).line lhs.line
     arg = remove_begin(arg)
     result = case asgn_op # REFACTOR
              when :"||" then
                lhs << arg
-               s(:op_asgn_or, self.gettable(name), lhs)
+               s(:op_asgn_or, name, lhs)
              when :"&&" then
                lhs << arg
-               s(:op_asgn_and, self.gettable(name), lhs)
+               s(:op_asgn_and, name, lhs)
              else
-               # TODO: why [2] ?
-               lhs[2] = new_call(self.gettable(name), asgn_op, argl(arg))
+               lhs << new_call(name, asgn_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
 
@@ -719,9 +1030,9 @@ module RubyParserStuff
     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 +1042,49 @@ module RubyParserStuff
     result
   end
 
+  def new_qsym_list
+    result = s(:array).line lexer.lineno
+    self.lexer.fixup_lineno
+    result
+  end
+
+  def new_qsym_list_entry val
+    _, str, _ = val
+    result = s(:lit, str.to_sym).line lexer.lineno
+    self.lexer.fixup_lineno
+    result
+  end
+
+  def new_qword_list
+    result = s(:array).line lexer.lineno
+    self.lexer.fixup_lineno
+    result
+  end
+
+  def new_qword_list_entry val
+    _, str, _ = val
+    str.force_encoding("ASCII-8BIT") unless str.valid_encoding?
+    result = s(:str, str).line lexer.lineno # TODO: problematic? grab from parser
+    self.lexer.fixup_lineno
+    result
+  end
+
   def new_regexp val
-    node = val[1] || s(:str, '')
-    options = val[2]
+    _, node, options = val
+
+    node ||= s(:str, "").line lexer.lineno
 
     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
@@ -778,7 +1117,7 @@ module RubyParserStuff
       end
       node << o if o and o != 0
     else
-      node = s(:dregx, '', node);
+      node = s(:dregx, "", node).line node.line
       node.sexp_type = :dregx_once if options =~ /o/
       node << o if o and o != 0
     end
@@ -786,19 +1125,20 @@ 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]
 
@@ -819,63 +1159,36 @@ module RubyParserStuff
   end
 
   def new_string val
-    str = val[0]
+    str, = val
     str.force_encoding("UTF-8")
+    # TODO: remove:
     str.force_encoding("ASCII-8BIT") unless str.valid_encoding?
-    result = s(:str, str)
+    result = s(:str, str).line lexer.lineno
     self.lexer.fixup_lineno str.count("\n")
     result
   end
 
-  def new_qword_list_entry val
-    str = val[1]
-    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
-  end
-
-  def new_qsym_list
-    result = s(:array)
-    self.lexer.fixup_lineno
-    result
-  end
-
-  def new_qsym_list_entry val
-    result = s(:lit, val[1].to_sym)
-    self.lexer.fixup_lineno
-    result
+  def new_super args
+    if args && args.node_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_symbol_list
-    result = s(:array)
+    result = s(:array).line lexer.lineno
     self.lexer.fixup_lineno
     result
   end
 
   def new_symbol_list_entry val
-    _list, sym, _nil = val # TODO: use _list
-    result = val[1]
+    _, sym, _ = val
+
+    sym ||= s(:str, "")
 
-    result ||= s(:str, "")
+    line = lexer.lineno
 
     case sym.sexp_type
     when :dstr then
@@ -883,26 +1196,21 @@ module RubyParserStuff
     when :str then
       sym = s(:lit, sym.last.to_sym)
     else
-      sym = s(:dsym, "", sym || s(:str, ""))
+      sym = s(:dsym, "", sym || s(:str, "").line(line))
     end
+
+    sym.line line
+
     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
 
@@ -935,20 +1243,36 @@ 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
+    result = s(:array).line lexer.lineno
+    self.lexer.fixup_lineno
+    result
+  end
+
+  def new_word_list_entry val
+    _, word, _ = val
+    result = word.sexp_type == :evstr ? s(:dstr, "", word).line(word.line) : word
+    self.lexer.fixup_lineno
+    result
+  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
@@ -957,113 +1281,32 @@ module RubyParserStuff
     raise SyntaxError, "Block argument should not be given." if
       args && args.node_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
+    args = s(:arglist, args).line args.line unless args.sexp_type == :arglist
 
-    return s(:yield, *args.sexp_body)
+    s(:yield, *args.sexp_body).line args.line
   end
 
   def next_token
     token = self.lexer.next_token
 
     if token and token.first != RubyLexer::EOF then
+      self.last_token_type = token
       return token
     else
-      return [false, '$end']
-    end
-  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
-    else
-      raise "unknown lhs #{lhs.inspect} w/ #{rhs.inspect}"
-    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
-
-    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[3..-1] if first =~ /\A\xEF\xBB\xBF/
-
-    encoding = $1.strip if header.find { |s|
-      s[/^#.*?-\*-.*?coding:\s*([^ ;]+).*?-\*-/, 1] ||
-      s[/^#.*(?:en)?coding(?:\s*[:=])\s*([\w-]+)/, 1]
-    }
-
-    if encoding then
-      if has_enc then
-        encoding.sub!(/utf-8-.+$/, '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
+      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)
+    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
 
   ##
@@ -1078,7 +1321,7 @@ module RubyParserStuff
 
       self.file = file.dup
 
-      @yydebug = ENV.has_key? 'DEBUG'
+      @yydebug = ENV.has_key? "DEBUG"
 
       # HACK -- need to get tests passing more than have graceful code
       self.lexer.ss = RPStringScanner.new str
@@ -1087,40 +1330,29 @@ module RubyParserStuff
     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 +1367,7 @@ 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 +1376,17 @@ 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 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
@@ -1173,17 +1404,34 @@ 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]
+    else
+      col
+    end
   end
 
+  alias remove_whitespace_width whitespace_width
+
   class Keyword
     include RubyLexer::State::Values
 
@@ -1214,8 +1462,10 @@ 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 +1495,7 @@ 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 +1575,6 @@ module RubyParserStuff
       @debug = debug
     end
 
-    def reset
-      @stack = [false]
-      log :reset if debug
-    end
-
     def inspect
       "StackState(#{@name}, #{@stack.inspect})"
     end
@@ -1366,16 +1611,21 @@ 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