module RubyParserStuff

Constants

ARG_TYPES
ENCODING_ORDER

This is in sorted order of occurrence according to charlock_holmes against 500k files, with UTF_8 forced to the top.

Overwrite this contstant if you need something different.

JUMP_TYPE
KEEP_COMMENT_TOKENS
TAB_WIDTH
VERSION

Attributes

canonicalize_conditions[RW]

Canonicalize conditionals. Eg:

not x ? a : b

becomes:

x ? b : a
env[R]
file[RW]
in_argdef[RW]
in_def[RW]
in_kwarg[RW]
in_single[RW]
last_token_type[RW]

The last token type returned from next_token

lexer[RW]

Public Class Methods

deprecate(old, new) click to toggle source
# File lib/ruby_parser_extras.rb, line 66
def self.deprecate old, new
  define_method old do |*args|
    warn "DEPRECATED: #{old} -> #{new} from #{caller.first}"
    send new, *args
  end
end
new(options = {}) click to toggle source
Calls superclass method
# File lib/ruby_parser_extras.rb, line 117
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

Public Instance Methods

_racc_do_reduce(arg, act) click to toggle source
Calls superclass method
# File lib/ruby_parser_extras.rb, line 81
def _racc_do_reduce arg, act
  x = super

  @racc_vstack.grep(Sexp).each do |sexp|
    sexp.check_line_numbers
  end
  x
end
arg_concat(node1, node2) click to toggle source
# File lib/ruby_parser_extras.rb, line 135
def arg_concat node1, node2 # TODO: nuke
  raise "huh" unless node2

  splat = s(:splat, node2)
  splat.line node2.line

  node1 << splat
end
argl(x) click to toggle source
# File lib/ruby_parser_extras.rb, line 144
def argl x
  x = s(:arglist, x) if x and x.sexp_type == :array
  x
end
args(args) click to toggle source
# File lib/ruby_parser_extras.rb, line 149
def args args
  result = s(:args)

  ss = args.grep Sexp
  if ss.empty? then
    result.line lexer.lineno
  else
    result.line ss.first.line
    result.line_max = ss.first.line_max
  end

  args.each do |arg|
    if arg.instance_of? Array and arg.size == 2 and arg.last.is_a? Numeric then
      arg = arg.first
    end

    case arg
    when Sexp then
      case arg.sexp_type
      when :args, :block, :array, :call_args then # HACK call_args mismatch
        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
        name = arg.last
        self.env[name] = :lvar
        if Sexp === result.last and result.last.sexp_type == :shadow then
          result.last << name
        else
          result << arg
        end
      when :masgn, :block_pass, :hash then # HACK: remove. prolly call_args
        result << arg
      else
        raise "unhandled: #{arg.sexp_type} in #{args.inspect}"
      end
    when Symbol then
      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
      raise "unhandled: #{arg.inspect} in #{args.inspect}"
    end
  end

  result
end
array_pat_concat(lhs, rhs) click to toggle source
# File lib/ruby_parser_extras.rb, line 756
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
array_to_hash(array) click to toggle source
# File lib/ruby_parser_extras.rb, line 238
def array_to_hash array
  case array.sexp_type
  when :kwsplat then
    array
  else
    s(:hash, *array.sexp_body).line array.line
  end
end
ary_to_pat(ary) click to toggle source
# File lib/ruby_parser_extras.rb, line 736
def ary_to_pat ary
  pat = ary.dup
  pat.sexp_type = :array_TAIL

  new_array_pattern nil, nil, pat, ary.line
end
aryset(receiver, index) click to toggle source
# File lib/ruby_parser_extras.rb, line 247
def aryset receiver, index
  index ||= s()
  l = receiver.line
  result = s(:attrasgn, receiver, :"[]=",
             *index.sexp_body).compact # [].sexp_body => nil
  result.line = l
  result
end
assignable(lhs, value = nil) click to toggle source
# File lib/ruby_parser_extras.rb, line 256
def assignable(lhs, value = nil)
  id, line = lhs
  id = id.to_sym

  result = case id
           when /^@@/ then
             asgn = in_def || in_single > 0
             s((asgn ? :cvasgn : :cvdecl), id)
           when /^@/ then
             s(:iasgn, id)
           when /^\$/ then
             s(:gasgn, id)
           when /^[A-Z]/ then
             s(:cdecl, id)
           else
             case self.env[id]
             when :lvar, :dvar, nil then
               s(:lasgn, id)
             else
               raise "wtf? unknown type: #{self.env[id]}"
             end
           end

  self.env[id] ||= :lvar if result.sexp_type == :lasgn

  result << value if value
  result.line line
  result
end
attrset_id?(id) click to toggle source
# File lib/ruby_parser_extras.rb, line 223
def attrset_id? id
  id =~ /^\[\]=$|^\w+=$/
end
backref_assign_error(ref) click to toggle source
# File lib/ruby_parser_extras.rb, line 286
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
block_append(head, tail) click to toggle source
# File lib/ruby_parser_extras.rb, line 300
def block_append(head, tail)
  return head if tail.nil?
  return tail if head.nil?

  line = [head.line, tail.line].compact.min

  head = remove_begin(head)
  head = s(:block, head).line(line) unless head.sexp_type == :block

  # head.line = line
  head << tail
end
block_dup_check(call_or_args, block) click to toggle source
# File lib/ruby_parser_extras.rb, line 313
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
block_var(*args) click to toggle source
# File lib/ruby_parser_extras.rb, line 318
def block_var *args
  result = self.args args
  result.sexp_type = :masgn
  result
end
call_args(args) click to toggle source
# File lib/ruby_parser_extras.rb, line 324
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
clean_mlhs(sexp) click to toggle source
# File lib/ruby_parser_extras.rb, line 361
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
cond(node) click to toggle source
# File lib/ruby_parser_extras.rb, line 382
def cond node
  return nil if node.nil?
  node = value_expr node

  case node.sexp_type
  when :lit then
    if Regexp === node.last then
      s(:match, node)
    else
      node
    end
  when :and then
    _, lhs, rhs = node
    s(:and,  cond(lhs), cond(rhs))
  when :or then
    _, lhs, rhs = node
    s(:or,  cond(lhs), cond(rhs))
  when :dot2 then
    label = "flip#{node.hash}"
    env[label] = :lvar
    _, lhs, rhs = node
    s(:flip2, lhs, rhs) # TODO: recurse?
  when :dot3 then
    label = "flip#{node.hash}"
    env[label] = :lvar
    _, lhs, rhs = node
    s(:flip3, lhs, rhs)
  else
    node
  end.line node.line
end
debug(n) click to toggle source
# File lib/ruby_parser_extras.rb, line 1633
def debug n
  if ENV["PRY"] then
    require "pry"; binding.pry
  end

  raise RubyParser::SyntaxError, "debug #{n}"
end
debug20(n, v = nil, r = nil) click to toggle source
# File lib/ruby_parser_extras.rb, line 62
def debug20 n, v = nil, r = nil
  raise "not yet #{n} #{v.inspect} => #{r.inspect}" unless $good20[n]
end
dedent(sexp) click to toggle source
# File lib/ruby_parser_extras.rb, line 414
def dedent sexp
  dedent_count = dedent_size sexp

  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

        s(:str, str).line obj.line
      else
        warn "unprocessed sexp %p" % [obj]
      end
    else
      warn "unprocessed: %p" % [obj]
    end
  }
end
dedent_size(sexp) click to toggle source
# File lib/ruby_parser_extras.rb, line 449
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
dedent_string(string, width) click to toggle source
# File lib/ruby_parser_extras.rb, line 479
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
  string[characters_skipped..-1]
end
do_parse() click to toggle source

for pure ruby systems only

# File lib/ruby_parser_extras.rb, line 76
def do_parse
  _racc_do_parse_rb(_racc_setup, false)
end
end_args(args) click to toggle source
# File lib/ruby_parser_extras.rb, line 217
def end_args args
  lexer.lex_state     = RubyLexer::State::Values::EXPR_BEG
  lexer.command_start = true
  self.args args
end
endless_method_name(defn_or_defs) click to toggle source
# File lib/ruby_parser_extras.rb, line 227
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
gettable(id) click to toggle source
# File lib/ruby_parser_extras.rb, line 498
def gettable(id)
  id = id.to_sym if String === id

  result = case id.to_s
           when /^@@/ then
             s(:cvar, id)
           when /^@/ then
             s(:ivar, id)
           when /^\$/ then
             s(:gvar, id)
           when /^[A-Z]/ then
             s(:const, id)
           else
             type = env[id]
             if type then
               s(type, id)
             else
               new_call(nil, id)
             end
           end

  raise "identifier #{id.inspect} is not valid" unless result

  result
end
hack_encoding(str, extra = nil) click to toggle source
# File lib/ruby_parser_extras.rb, line 524
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
handle_encoding(str) click to toggle source

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.

# File lib/ruby_parser_extras.rb, line 568
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/

  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
  end

  str
end
invert_block_call(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 599
def invert_block_call val
  ret, iter = val
  type, call = ret

  iter.insert 1, call

  ret = s(type).line ret.line

  [iter, ret]
end
inverted?(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 610
def inverted? val
  JUMP_TYPE[val[0].sexp_type]
end
list_append(list, item) click to toggle source
# File lib/ruby_parser_extras.rb, line 614
def list_append list, item # TODO: nuke me *sigh*
  return s(:array, item) unless list
  list = s(:array, list) unless Sexp === list && list.sexp_type == :array
  list << item
end
list_prepend(item, list) click to toggle source
# File lib/ruby_parser_extras.rb, line 620
def list_prepend item, list # TODO: nuke me *sigh*
  list = s(:array, list) unless Sexp === list && list.sexp_type == :array
  list.insert 1, item
  list
end
literal_concat(head, tail) click to toggle source
# File lib/ruby_parser_extras.rb, line 626
def literal_concat head, tail # TODO: ugh. rewrite
  return tail unless head
  return head unless tail

  htype, ttype = head.sexp_type, tail.sexp_type

  head = s(:dstr, "", head).line head.line if htype == :evstr

  case ttype
  when :str then
    if htype == :str
      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
      head << tail
    end
  when :dstr then
    if htype == :str then
      lineno = head.line
      tail[1] = head.last + tail[1]
      head = tail
      head.line = lineno
    else
      tail.sexp_type = :array
      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.file = f
      head.line = l
    end

    if head.size == 2 and tail.size > 1 and tail[1].sexp_type == :str then
      head[-1] = head[-1].dup if head[-1].frozen?
      head.last << tail[1].last
      head.sexp_type = :str if head.size == 2 # HACK ?
    else
      head.push(tail)
    end
  else
    x = [head, tail]
    raise "unknown type: #{x.inspect}"
  end

  return head
end
local_pop(in_def) click to toggle source
# File lib/ruby_parser_extras.rb, line 681
def local_pop in_def
  lexer.cond.pop # group = local_pop
  lexer.cmdarg.pop
  self.env.unextend
  self.in_def = in_def
end
logical_op(type, left, right) click to toggle source
# File lib/ruby_parser_extras.rb, line 688
def logical_op type, left, right
  left = value_expr left

  if left and left.sexp_type == type and not left.paren then
    node, rhs = left, nil

    loop do
      _, _lhs, rhs = node
      break unless rhs && rhs.sexp_type == type and not rhs.paren
      node = rhs
    end

    node.pop
    node << s(type, rhs, right).line(rhs.line)

    return left
  end

  result = s(type, left, right)
  result.line left.line if left.line
  result
end
new_aref(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 711
def new_aref val
  val[2] ||= s(:arglist)
  val[2].sexp_type = :arglist if val[2].sexp_type == :array # REFACTOR
  new_call val[0], :"[]", val[2]
end
new_arg(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 717
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
new_array_pattern(const, pre_arg, arypat, loc) click to toggle source
# File lib/ruby_parser_extras.rb, line 743
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
new_array_pattern_tail(pre_args, has_rest, rest_arg, post_args) click to toggle source
# File lib/ruby_parser_extras.rb, line 772
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
new_assign(lhs, rhs) click to toggle source
# File lib/ruby_parser_extras.rb, line 791
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
new_attrasgn(recv, meth, call_op = :".") click to toggle source
# File lib/ruby_parser_extras.rb, line 810
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
new_begin(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 828
def new_begin val
  (_, line), _, body, _ = val

  result = body ? s(:begin, body) : s(:nil)
  result.line line

  result
end
new_body(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 837
def new_body val
  body, resbody, elsebody, ensurebody = val

  result = body

  if resbody then
    result = s(:rescue)
    result << body if body

    res = resbody

    while res do
      result << res
      res = res.resbody(true)
    end

    result << elsebody if elsebody

    result.line = (body || resbody).line
  end

  if elsebody and not resbody then
    warning("else without rescue is useless")
    result = s(:begin, result).line result.line if result
    result = block_append(result, elsebody)
  end

  if ensurebody
    lineno = (result || ensurebody).line
    result = s(:ensure, result, ensurebody).compact.line lineno
  end

  result
end
new_brace_body(args, body, lineno) click to toggle source
# File lib/ruby_parser_extras.rb, line 872
def new_brace_body args, body, lineno
  new_iter(nil, args, body).line lineno
end
new_call(recv, meth, args = nil, call_op = :".") click to toggle source
# File lib/ruby_parser_extras.rb, line 876
def new_call recv, meth, args = nil, call_op = :"."
  call_op = call_op.first if Array === call_op

  result = case call_op.to_sym
           when :"."
             s(:call, recv, meth)
           when :"&."
             s(:safe_call, recv, meth)
           else
             raise "unknown call operator: `#{type.inspect}`"
           end

  # TODO: need a test with f(&b) to produce block_pass
  # TODO: need a test with f(&b) { } to produce warning

  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 = recv.line if recv
  result.line ||= lexer.lineno

  result
end
new_case(expr, body, line) click to toggle source
# File lib/ruby_parser_extras.rb, line 911
def new_case expr, body, line
  result = s(:case, expr)

  while body and [:when, :in].include? body.sexp_type
    result << body
    body = body.delete_at 3
  end

  result[2..-1].each do |node|
    block = node.block(:delete)
    node.concat block.sexp_body if block
  end

  # else
  body = nil if body == s(:block)
  result << body

  result.line = line
  result
end
new_class(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 932
def new_class val
  (_, line, comment), path, superclass, _, body, (_, line_max) = val

  path = path.first if path.instance_of? Array

  result = s(:class, path, superclass)

  if body then
    if body.sexp_type == :block then
      result.push(*body.sexp_body)
    else
      result.push body
    end
  end

  result.line = line
  result.line_max = line_max
  result.comments = comment if comment
  result
end
new_compstmt(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 953
def new_compstmt val
  result = void_stmts(val.grep(Sexp)[0])
  result = remove_begin(result) if result
  result
end
new_const_op_asgn(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 959
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
new_defn(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 976
def new_defn val
  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

  body ||= s(:nil).line line

  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
new_defs(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 1047
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
new_do_body(args, body, lineno) click to toggle source
# File lib/ruby_parser_extras.rb, line 1077
def new_do_body args, body, lineno
  new_iter(nil, args, body).line(lineno)
end
new_endless_defn(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 1001
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
      s(:defn, name, args, body).line line
    end

  local_pop in_def
  endless_method_name result

  result.comments = comment if comment

  result
end
new_endless_defs(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 1023
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

  result =
    if resbody then
      s(:defs, recv, name, args,
        new_rescue(body,
                   new_resbody(s(:array).line(line),
                               resbody))).line line
    else
      s(:defs, recv, name, args, body).line(line)
    end

  self.in_single -= 1
  local_pop in_def
  endless_method_name result

  result.comments = comment if comment

  result
end
new_find_pattern(const, pat) click to toggle source
# File lib/ruby_parser_extras.rb, line 1081
def new_find_pattern const, pat
  pat.sexp_type = :find_pat
  pat.insert 1, const
end
new_find_pattern_tail(lhs, mid, rhs) click to toggle source
# File lib/ruby_parser_extras.rb, line 1086
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
new_for(expr, var, body) click to toggle source
# File lib/ruby_parser_extras.rb, line 1100
def new_for expr, var, body
  result = s(:for, expr, var).line(var.line)
  result << body if body
  result
end
new_hash(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 1106
def new_hash val
  _, line, assocs = val

  s(:hash).line(line).concat assocs.sexp_body
end
new_hash_pattern(const, hash_pat, loc) click to toggle source
# File lib/ruby_parser_extras.rb, line 1112
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
new_hash_pattern_tail(kw_args, kw_rest_arg, line) click to toggle source
# File lib/ruby_parser_extras.rb, line 1124
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
new_if(c, t, f) click to toggle source
# File lib/ruby_parser_extras.rb, line 1152
def new_if c, t, f
  l = [c.line, t && t.line, f && f.line].compact.min
  c = cond c
  c, t, f = c.last, f, t if c.sexp_type == :not and canonicalize_conditions
  s(:if, c, t, f).line(l)
end
new_in(pat, body, cases, line) click to toggle source
# File lib/ruby_parser_extras.rb, line 907
def new_in pat, body, cases, line
  s(:in, pat, body, cases).line line
end
new_iter(call, args, body) click to toggle source
# File lib/ruby_parser_extras.rb, line 1159
def new_iter call, args, body
  body ||= nil

  args ||= s(:args)
  args = s(:args, args) if Symbol === args

  result = s(:iter)
  result << call if call
  result << args
  result << body if body

  result.line call.line if call

  unless args == 0 then
    args.line call.line if call
    args.sexp_type = :args
  end

  result
end
new_masgn(lhs, rhs, wrap = false) click to toggle source
# File lib/ruby_parser_extras.rb, line 1180
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
end
new_masgn_arg(rhs, wrap = false) click to toggle source
# File lib/ruby_parser_extras.rb, line 1194
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
new_match(lhs, rhs) click to toggle source
# File lib/ruby_parser_extras.rb, line 1201
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
new_module(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 1225
def new_module val
  (_, line_min, comment), path, _, body, (_, line_max) = val

  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
      result.push(*body.sexp_body)
    else
      result.push body
    end
  end

  result.comments = comment if comment
  result
end
new_op_asgn(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 1245
def new_op_asgn val
  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, name, lhs).line lhs.line
           when :"&&" then
             lhs << arg
             s(:op_asgn_and, name, lhs).line lhs.line
           else
             lhs << new_call(name, op, argl(arg))
             lhs
           end

  result
end
new_op_asgn1(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 1266
def new_op_asgn1 val
  lhs, _, args, _, (op, _), rhs = val

  args.sexp_type = :arglist if args

  result = s(:op_asgn1, lhs, args, op.to_sym, rhs)
  result.line lhs.line
  result
end
new_op_asgn2(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 1276
def new_op_asgn2 val
  recv, (call_op, _), (meth, _), (op, _), arg = val
  meth = :"#{meth}="

  result = case call_op.to_sym
           when :"."
             s(:op_asgn2, recv, meth, op.to_sym, arg)
           when :"&."
             s(:safe_op_asgn2, recv, meth, op.to_sym, arg)
           else
             raise "unknown call operator: `#{type.inspect}`"
           end

  result.line = recv.line
  result
end
new_qsym_list() click to toggle source
# File lib/ruby_parser_extras.rb, line 1293
def new_qsym_list
  s(:array).line lexer.lineno
end
new_qsym_list_entry(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 1297
def new_qsym_list_entry val
  _, (str, line), _ = val
  s(:lit, str.to_sym).line line
end
new_qword_list() click to toggle source
# File lib/ruby_parser_extras.rb, line 1302
def new_qword_list
  s(:array).line lexer.lineno
end
new_qword_list_entry(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 1306
def new_qword_list_entry val
  _, (str, line), _ = val
  str.force_encoding("ASCII-8BIT") unless str.valid_encoding?
  s(:str, str).line line
end
new_regexp(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 1312
def new_regexp val
  (_, 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,
    }[c]
    raise "unknown regexp option: #{c}" unless v
    o += v
  end

  case node.sexp_type
  when :str then
    node.sexp_type = :lit
    node[1] = if k then
                Regexp.new(node[1], o, k)
              else
                begin
                  Regexp.new(node[1], o)
                rescue RegexpError => e
                  warn "WARNING: #{e.message} for #{node[1].inspect} #{options.inspect}"
                  begin
                    warn "WARNING: trying to recover with ENC_UTF8"
                    Regexp.new(node[1], Regexp::ENC_UTF8)
                  rescue RegexpError => e
                    warn "WARNING: trying to recover with ENC_NONE"
                    Regexp.new(node[1], Regexp::ENC_NONE)
                  end
                end
              end
  when :dstr then
    if options =~ /o/ then
      node.sexp_type = :dregx_once
    else
      node.sexp_type = :dregx
    end
    node << o if o and o != 0
  else
    node = s(:dregx, "", node).line line
    node.sexp_type = :dregx_once if options =~ /o/
    node << o if o and o != 0
  end

  node
end
new_resbody(cond, body) click to toggle source
# File lib/ruby_parser_extras.rb, line 1369
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
new_rescue(body, resbody) click to toggle source
# File lib/ruby_parser_extras.rb, line 1379
def new_rescue body, resbody
  s(:rescue, body, resbody).line body.line
end
new_sclass(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 1383
def new_sclass val
  (_, line), _, recv, in_def, _, in_single, body, _ = val

  result = s(:sclass, recv)

  if body then
    if body.sexp_type == :block then
      result.push(*body.sexp_body)
    else
      result.push body
    end
  end

  result.line = line
  self.in_def = in_def
  self.in_single = in_single
  result
end
new_string(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 1402
def new_string val
  (str, line), = val

  str.force_encoding("UTF-8")
  # TODO: remove:
  str.force_encoding("ASCII-8BIT") unless str.valid_encoding?
  s(:str, str).line line
end
new_super(args) click to toggle source
# File lib/ruby_parser_extras.rb, line 1411
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
new_symbol(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 1420
def new_symbol val
  name = val.last
  s(:lit, name.to_sym).line lexer.lineno
end
new_symbol_list() click to toggle source
# File lib/ruby_parser_extras.rb, line 1425
def new_symbol_list
  # TODO: hunt down and try to remove ALL lexer.lineno usage!
  s(:array).line lexer.lineno
end
new_symbol_list_entry(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 1430
def new_symbol_list_entry val
  _, sym, _ = val

  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).line sym.line
  else
    sym = s(:dsym, "", sym).line sym.line
  end

  sym
end
new_undef(n, m = nil) click to toggle source
# File lib/ruby_parser_extras.rb, line 1447
def new_undef n, m = nil
  if m then
    block_append(n, s(:undef, m).line(m.line))
  else
    s(:undef, n).line n.line
  end
end
new_until(block, expr, pre) click to toggle source
# File lib/ruby_parser_extras.rb, line 1455
def new_until block, expr, pre
  new_until_or_while :until, block, expr, pre
end
new_until_or_while(type, block, expr, pre) click to toggle source
# File lib/ruby_parser_extras.rb, line 1459
def new_until_or_while type, block, expr, pre
  other = type == :until ? :while : :until
  line = [block && block.line, expr.line].compact.min
  block, pre = block.last, false if block && block.sexp_type == :begin

  expr = cond expr

  result = unless expr.sexp_type == :not and canonicalize_conditions then
             s(type,  expr,      block, pre)
           else
             s(other, expr.last, block, pre)
           end

  result.line = line
  result
end
new_when(cond, body) click to toggle source
# File lib/ruby_parser_extras.rb, line 1476
def new_when cond, body
  s(:when, cond, body)
end
new_while(block, expr, pre) click to toggle source
# File lib/ruby_parser_extras.rb, line 1480
def new_while block, expr, pre
  new_until_or_while :while, block, expr, pre
end
new_word_list() click to toggle source
# File lib/ruby_parser_extras.rb, line 1484
def new_word_list
  s(:array).line lexer.lineno
end
new_word_list_entry(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 1488
def new_word_list_entry val
  _, word, _ = val
  word.sexp_type == :evstr ? s(:dstr, "", word).line(word.line) : word
end
new_xstring(val) click to toggle source
# File lib/ruby_parser_extras.rb, line 1493
def new_xstring val
  _, node = val

  node ||= s(:str, "").line lexer.lineno

  if node then
    case node.sexp_type
    when :str
      node.sexp_type = :xstr
    when :dstr
      node.sexp_type = :dxstr
    else
      node = s(:dxstr, "", node).line node.line
    end
  end

  node
end
new_yield(args = nil) click to toggle source
# File lib/ruby_parser_extras.rb, line 1512
def new_yield args = nil
  # TODO: raise args.inspect unless [:arglist].include? args.first # HACK
  raise "write a test 4" if args && args.sexp_type == :block_pass
  raise SyntaxError, "Block argument should not be given." if
    args && args.sexp_type == :block_pass

  args ||= s(:arglist).line lexer.lineno

  args.sexp_type = :arglist if [:call_args, :array].include? args.sexp_type
  args = s(:arglist, args).line args.line unless args.sexp_type == :arglist

  s(:yield, *args.sexp_body).line args.line
end
next_token() click to toggle source
# File lib/ruby_parser_extras.rb, line 1537
def next_token
  token = self.lexer.next_token

  if token and token.first != RubyLexer::EOF then
    self.last_token_type = token

    self.lexer.comment = nil unless KEEP_COMMENT_TOKENS.include? token.first

    return token
  elsif !token
    return self.lexer.next_token
  else
    return [false, false]
  end
end
on_error(et, ev, values) click to toggle source
Calls superclass method
# File lib/ruby_parser_extras.rb, line 1553
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
parse(str, file = "(string)", time = 10)
Alias for: process
pop_pktbl() click to toggle source
# File lib/ruby_parser_extras.rb, line 1143
def pop_pktbl
end
pop_pvtbl() click to toggle source
# File lib/ruby_parser_extras.rb, line 1149
def pop_pvtbl
end
prev_value_to_lineno(v) click to toggle source
# File lib/ruby_parser_extras.rb, line 1526
def prev_value_to_lineno v
  s, n = v
  if String === s then
    n
  else
    lexer.lineno
  end
end
process(str, file = "(string)", time = 10) click to toggle source

Parse str at path file and return a sexp. Raises Timeout::Error if it runs for more than time seconds.

# File lib/ruby_parser_extras.rb, line 1567
def process(str, file = "(string)", time = 10)
  str.freeze

  Timeout.timeout time do
    raise "bad val: #{str.inspect}" unless String === str

    self.lexer.string = handle_encoding str

    self.file = file.dup

    @yydebug = ENV.has_key? "DEBUG"

    do_parse
  end
end
Also aliased as: parse
push_pktbl() click to toggle source
# File lib/ruby_parser_extras.rb, line 1140
def push_pktbl
end
push_pvtbl() click to toggle source
# File lib/ruby_parser_extras.rb, line 1146
def push_pvtbl
end
remove_begin(node) click to toggle source
# File lib/ruby_parser_extras.rb, line 1585
def remove_begin node
  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
Also aliased as: value_expr
remove_whitespace_width(line, remove_width = nil)
Alias for: whitespace_width
reset() click to toggle source
# File lib/ruby_parser_extras.rb, line 1599
def reset
  lexer.reset
  self.in_def = false
  self.in_single = 0
  self.env.reset
  self.last_token_type = nil
end
ret_args(node) click to toggle source
# File lib/ruby_parser_extras.rb, line 1607
def ret_args node
  if node then
    raise "write a test 5" if node.sexp_type == :block_pass

    raise SyntaxError, "block argument should not be given" if
      node.sexp_type == :block_pass

    node.sexp_type = :array if node.sexp_type == :call_args
    node = node.last if node.sexp_type == :array && node.size == 2

    # HACK matz wraps ONE of the FOUR splats in a newline to
    # distinguish. I use paren for now. ugh
    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

  node
end
s(*args) click to toggle source
# File lib/ruby_parser_extras.rb, line 1626
def s(*args)
  result = Sexp.new(*args)
  # result.line ||= lexer.lineno if lexer.ss unless ENV["CHECK_LINE_NUMS"] # otherwise...
  result.file = self.file
  result
end
syntax_error(msg) click to toggle source
# File lib/ruby_parser_extras.rb, line 1641
def syntax_error msg
  raise RubyParser::SyntaxError, msg
end
Also aliased as: yyerror
value_expr(node)
Alias for: remove_begin
void_stmts(node) click to toggle source
# File lib/ruby_parser_extras.rb, line 1647
def void_stmts node
  return nil unless node
  return node unless node.sexp_type == :block

  if node.respond_to? :sexp_body= then
    node.sexp_body = node.sexp_body.map { |n| remove_begin n }
  else
    node[1..-1] = node[1..-1].map { |n| remove_begin(n) }
  end

  node
end
warning(s) click to toggle source
# File lib/ruby_parser_extras.rb, line 1660
def warning s
  # do nothing for now
end
whitespace_width(line, remove_width = nil) click to toggle source
# File lib/ruby_parser_extras.rb, line 1664
def whitespace_width line, remove_width = nil
  col = 0
  idx = 0

  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
Also aliased as: remove_whitespace_width
wrap(type, node) click to toggle source
# File lib/ruby_parser_extras.rb, line 1694
def wrap type, node
  value, line = node
  value = value.to_sym if value.respond_to? :to_sym
  s(type, value).line line
end
yyerror(msg)
Alias for: syntax_error