test_regexp.rb

# coding: US-ASCII
# frozen_string_literal: false
require 'test/unit'

class TestRegexp < Test::Unit::TestCase
  def setup
    @verbose = $VERBOSE
  end

  def teardown
    $VERBOSE = @verbose
  end

  def test_has_NOENCODING
    assert Regexp::NOENCODING
    re = //n
    assert_equal Regexp::NOENCODING, re.options
  end

  def test_ruby_dev_999
    assert_match(/(?<=a).*b/, "aab")
    assert_match(/(?<=\u3042).*b/, "\u3042ab")
  end

  def test_ruby_core_27247
    assert_match(/(a){2}z/, "aaz")
  end

  def test_ruby_dev_24643
    assert_nothing_raised("[ruby-dev:24643]") {
      /(?:(?:[a]*[a])?b)*a*$/ =~ "aabaaca"
    }
  end

  def test_ruby_talk_116455
    assert_match(/^(\w{2,}).* ([A-Za-z\xa2\xc0-\xff]{2,}?)$/n, "Hallo Welt")
  end

  def test_ruby_dev_24887
    assert_equal("a".gsub(/a\Z/, ""), "")
  end

  def test_ruby_dev_31309
    assert_equal('Ruby', 'Ruby'.sub(/[^a-z]/i, '-'))
  end

  def test_premature_end_char_property
    ["\\p{",
     "\\p{".dup.force_encoding("UTF-8"),
     "\\p{".dup.force_encoding("US-ASCII")
    ].each do |string|
      assert_raise(RegexpError) do
        Regexp.new(string)
      end
    end
  end

  def test_assert_normal_exit
    # moved from knownbug.  It caused core.
    Regexp.union("a", "a")
  end

  def test_to_s
    assert_equal '(?-mix:\x00)', Regexp.new("\0").to_s

    str = "abcd\u3042"
    [:UTF_16BE, :UTF_16LE, :UTF_32BE, :UTF_32LE].each do |es|
      enc = Encoding.const_get(es)
      rs = Regexp.new(str.encode(enc)).to_s
      assert_equal("(?-mix:abcd\u3042)".encode(enc), rs)
      assert_equal(enc, rs.encoding)
    end
  end

  def test_to_s_extended_subexp
    re = /#\g#{"\n"}/x
    re = /#{re}/
    assert_warn('', '[ruby-core:82328] [Bug #13798]') {re.to_s}
  end

  def test_extended_comment_invalid_escape_bug_18294
    assert_separately([], <<-RUBY)
      re = / C:\\\\[a-z]{5} # e.g. C:\\users /x
      assert_match(re, 'C:\\users')
      assert_not_match(re, 'C:\\user')

      re = /
        foo  # \\M-ca
        bar
      /x
      assert_match(re, 'foobar')
      assert_not_match(re, 'foobaz')

      re = /
        f[#o]o  # \\M-ca
        bar
      /x
      assert_match(re, 'foobar')
      assert_not_match(re, 'foobaz')

      re = /
        f[[:alnum:]#]o  # \\M-ca
        bar
      /x
      assert_match(re, 'foobar')
      assert_not_match(re, 'foobaz')

      re = /
        f(?# \\M-ca)oo  # \\M-ca
        bar
      /x
      assert_match(re, 'foobar')
      assert_not_match(re, 'foobaz')

      re = /f(?# \\M-ca)oobar/
      assert_match(re, 'foobar')
      assert_not_match(re, 'foobaz')

      re = /[-(?# fca)]oobar/
      assert_match(re, 'foobar')
      assert_not_match(re, 'foobaz')

      re = /f(?# ca\0\\M-ca)oobar/
      assert_match(re, 'foobar')
      assert_not_match(re, 'foobaz')
    RUBY

    assert_raise(SyntaxError) {eval "/\\users/x"}
    assert_raise(SyntaxError) {eval "/[\\users]/x"}
    assert_raise(SyntaxError) {eval "/(?<\\users)/x"}
    assert_raise(SyntaxError) {eval "/# \\users/"}
  end

  def test_nonextended_section_of_extended_regexp_bug_19379
    assert_separately([], <<-'RUBY')
      re = /(?-x:#)/x
      assert_match(re, '#')
      assert_not_match(re, '-')

      re = /(?xi:#
      y)/
      assert_match(re, 'Y')
      assert_not_match(re, '-')

      re = /(?mix:#
      y)/
      assert_match(re, 'Y')
      assert_not_match(re, '-')

      re = /(?x-im:#
      y)/i
      assert_match(re, 'y')
      assert_not_match(re, 'Y')

      re = /(?-imx:(?xim:#
      y))/x
      assert_match(re, 'y')
      assert_not_match(re, '-')

      re = /(?x)#
      y/
      assert_match(re, 'y')
      assert_not_match(re, 'Y')

      re = /(?mx-i)#
      y/i
      assert_match(re, 'y')
      assert_not_match(re, 'Y')

      re = /(?-imx:(?xim:#
      (?-x)y#))/x
      assert_match(re, 'Y#')
      assert_not_match(re, '-#')

      re = /(?imx:#
      (?-xim:#(?im)#(?x)#
      )#
      (?x)#
      y)/
      assert_match(re, '###Y')
      assert_not_match(re, '###-')

      re = %r{#c-\w+/comment/[\w-]+}
      re = %r{https?://[^/]+#{re}}x
      assert_match(re, 'http://foo#c-x/comment/bar')
      assert_not_match(re, 'http://foo#cx/comment/bar')
    RUBY
  end

  def test_utf8_comment_in_usascii_extended_regexp_bug_19455
    assert_separately([], <<-RUBY)
      assert_equal(Encoding::UTF_8, /(?#\u1000)/x.encoding)
      assert_equal(Encoding::UTF_8, /#\u1000/x.encoding)
    RUBY
  end

  def test_union
    assert_equal :ok, begin
      Regexp.union(
        "a",
        Regexp.new("\xc2\xa1".force_encoding("euc-jp")),
        Regexp.new("\xc2\xa1".force_encoding("utf-8")))
      :ng
    rescue ArgumentError
      :ok
    end
    re = Regexp.union(/\//, "")
    re2 = eval(re.inspect)
    assert_equal(re.to_s, re2.to_s)
    assert_equal(re.source, re2.source)
    assert_equal(re, re2)
  end

  def test_word_boundary
    assert_match(/\u3042\b /, "\u3042 ")
    assert_not_match(/\u3042\ba/, "\u3042a")
  end

  def test_named_capture
    m = /&(?<foo>.*?);/.match("aaa &amp; yyy")
    assert_equal("amp", m["foo"])
    assert_equal("amp", m[:foo])
    assert_equal(5, m.begin(:foo))
    assert_equal(8, m.end(:foo))
    assert_equal([5,8], m.offset(:foo))

    assert_equal("aaa [amp] yyy",
      "aaa &amp; yyy".sub(/&(?<foo>.*?);/, '[\k<foo>]'))

    assert_equal('#<MatchData "&amp; y" foo:"amp">',
      /&(?<foo>.*?); (y)/.match("aaa &amp; yyy").inspect)
    assert_equal('#<MatchData "&amp; y" 1:"amp" 2:"y">',
      /&(.*?); (y)/.match("aaa &amp; yyy").inspect)
    assert_equal('#<MatchData "&amp; y" foo:"amp" bar:"y">',
      /&(?<foo>.*?); (?<bar>y)/.match("aaa &amp; yyy").inspect)
    assert_equal('#<MatchData "&amp; y" foo:"amp" foo:"y">',
      /&(?<foo>.*?); (?<foo>y)/.match("aaa &amp; yyy").inspect)

    /(?<_id>[A-Za-z_]+)/ =~ "!abc"
    assert_not_nil(Regexp.last_match)
    assert_equal("abc", Regexp.last_match(1))
    assert_equal("abc", Regexp.last_match(:_id))

    /a/ =~ "b" # doesn't match.
    assert_equal(nil, Regexp.last_match)
    assert_equal(nil, Regexp.last_match(1))
    assert_equal(nil, Regexp.last_match(:foo))

    bug11825_name = "\u{5b9d 77f3}"
    bug11825_str = "\u{30eb 30d3 30fc}"
    bug11825_re = /(?<#{bug11825_name}>)#{bug11825_str}/

    assert_equal(["foo", "bar"], /(?<foo>.)(?<bar>.)/.names)
    assert_equal(["foo"], /(?<foo>.)(?<foo>.)/.names)
    assert_equal([], /(.)(.)/.names)
    assert_equal([bug11825_name], bug11825_re.names)

    assert_equal(["foo", "bar"], /(?<foo>.)(?<bar>.)/.match("ab").names)
    assert_equal(["foo"], /(?<foo>.)(?<foo>.)/.match("ab").names)
    assert_equal([], /(.)(.)/.match("ab").names)
    assert_equal([bug11825_name], bug11825_re.match(bug11825_str).names)

    assert_equal({"foo"=>[1], "bar"=>[2]},
                 /(?<foo>.)(?<bar>.)/.named_captures)
    assert_equal({"foo"=>[1, 2]},
                 /(?<foo>.)(?<foo>.)/.named_captures)
    assert_equal({}, /(.)(.)/.named_captures)

    assert_equal("a[b]c", "abc".sub(/(?<x>[bc])/, "[\\k<x>]"))

    assert_equal("o", "foo"[/(?<bar>o)/, "bar"])
    assert_equal("o", "foo"[/(?<@bar>o)/, "@bar"])
    assert_equal("o", "foo"[/(?<@bar>.)\g<@bar>\k<@bar>/, "@bar"])

    s = "foo"
    s[/(?<bar>o)/, "bar"] = "baz"
    assert_equal("fbazo", s)

    /.*/ =~ "abc"
    "a".sub("a", "")
    assert_raise(IndexError) {Regexp.last_match(:_id)}
  end

  def test_named_capture_with_nul
    bug9902 = '[ruby-dev:48275] [Bug #9902]'

    m = /(?<a>.*)/.match("foo")
    assert_raise(IndexError, bug9902) {m["a\0foo"]}
    assert_raise(IndexError, bug9902) {m["a\0foo".to_sym]}

    m = Regexp.new("(?<foo\0bar>.*)").match("xxx")
    assert_raise(IndexError, bug9902) {m["foo"]}
    assert_raise(IndexError, bug9902) {m["foo".to_sym]}
    assert_nothing_raised(IndexError, bug9902) {
      assert_equal("xxx", m["foo\0bar"], bug9902)
      assert_equal("xxx", m["foo\0bar".to_sym], bug9902)
    }
  end

  def test_named_capture_nonascii
    bug9903 = '[ruby-dev:48278] [Bug #9903]'

    key = "\xb1\xb2".force_encoding(Encoding::EUC_JP)
    m = /(?<#{key}>.*)/.match("xxx")
    assert_equal("xxx", m[key])
    assert_raise(IndexError, bug9903) {m[key.dup.force_encoding(Encoding::Shift_JIS)]}
  end

  def test_match_data_named_captures
    assert_equal({'a' => '1', 'b' => '2', 'c' => nil}, /^(?<a>.)(?<b>.)(?<c>.)?/.match('12').named_captures)
    assert_equal({'a' => '1', 'b' => '2', 'c' => '3'}, /^(?<a>.)(?<b>.)(?<c>.)?/.match('123').named_captures)
    assert_equal({'a' => '1', 'b' => '2', 'c' => ''}, /^(?<a>.)(?<b>.)(?<c>.?)/.match('12').named_captures)

    assert_equal({a: '1', b: '2', c: ''}, /^(?<a>.)(?<b>.)(?<c>.?)/.match('12').named_captures(symbolize_names: true))
    assert_equal({'a' => '1', 'b' => '2', 'c' => ''}, /^(?<a>.)(?<b>.)(?<c>.?)/.match('12').named_captures(symbolize_names: false))

    assert_equal({'a' => 'x'}, /(?<a>x)|(?<a>y)/.match('x').named_captures)
    assert_equal({'a' => 'y'}, /(?<a>x)|(?<a>y)/.match('y').named_captures)

    assert_equal({'a' => '1', 'b' => '2'}, /^(.)(?<a>.)(?<b>.)/.match('012').named_captures)
    assert_equal({'a' => '2'}, /^(?<a>.)(?<a>.)/.match('12').named_captures)

    assert_equal({}, /^(.)/.match('123').named_captures)
  end

  def test_assign_named_capture
    assert_equal("a", eval('/(?<foo>.)/ =~ "a"; foo'))
    assert_equal(nil, eval('/(?<@foo>.)/ =~ "a"; defined?(@foo)'))
    assert_equal("a", eval('foo = 1; /(?<foo>.)/ =~ "a"; foo'))
    assert_equal("a", eval('1.times {|foo| /(?<foo>.)/ =~ "a"; break foo }'))
    assert_nothing_raised { eval('/(?<Foo>.)/ =~ "a"') }
    assert_nil(eval('/(?<Foo>.)/ =~ "a"; defined? Foo'))
  end

  def test_assign_named_capture_to_reserved_word
    /(?<nil>.)/ =~ "a"
    assert_not_include(local_variables, :nil, "[ruby-dev:32675]")

    def (obj = Object.new).test(s, nil: :ng)
      /(?<nil>.)/ =~ s
      binding.local_variable_get(:nil)
    end
    assert_equal("b", obj.test("b"))

    tap do |nil: :ng|
      /(?<nil>.)/ =~ "c"
      assert_equal("c", binding.local_variable_get(:nil))
    end
  end

  def test_assign_named_capture_to_const
    %W[C \u{1d402}].each do |name|
      assert_equal(:ok, Class.new.class_eval("#{name} = :ok; /(?<#{name}>.*)/ =~ 'ng'; #{name}"))
    end
  end

  def test_assign_named_capture_trace
    bug = '[ruby-core:79940] [Bug #13287]'
    assert_normal_exit("#{<<-"begin;"}\n#{<<-"end;"}", bug)
    begin;
      / (?<foo>.*)/ =~ "bar" &&
        true
    end;
  end

  def test_match_regexp
    r = /./
    m = r.match("a")
    assert_equal(r, m.regexp)
    re = /foo/
    assert_equal(re, re.match("foo").regexp)
  end

  def test_match_lambda_multithread
    bug17507 = "[ruby-core:101901]"
    str = "a-x-foo-bar-baz-z-b"

    worker = lambda do
      m = /foo-([A-Za-z0-9_\.]+)-baz/.match(str)
      assert_equal("bar", m[1], bug17507)

      # These two lines are needed to trigger the bug
      File.exist? "/tmp"
      str.gsub(/foo-bar-baz/, "foo-abc-baz")
    end

    def self. threaded_test(worker)
      6.times.map {Thread.new {10_000.times {worker.call}}}.each(&:join)
    end

    # The bug only occurs in a method calling a block/proc/lambda
    threaded_test(worker)
  end

  def test_source
    bug5484 = '[ruby-core:40364]'
    assert_equal('', //.source)
    assert_equal('\:', /\:/.source, bug5484)
    assert_equal(':', %r:\::.source, bug5484)
  end

  def test_source_escaped
    expected, result = "$*+.?^|".each_char.map {|c|
      [
        ["\\#{c}", "\\#{c}", 1],
        begin
          re = eval("%r#{c}\\#{c}#{c}", nil, __FILE__, __LINE__)
          t = eval("/\\#{c}/", nil, __FILE__, __LINE__).source
        rescue SyntaxError => e
          [e, t, nil]
        else
          [re.source, t, re =~ "a#{c}a"]
        end
      ]
    }.transpose
    assert_equal(expected, result)
  end

  def test_source_escaped_paren
    bug7610 = '[ruby-core:51088] [Bug #7610]'
    bug8133 = '[ruby-core:53578] [Bug #8133]'
    [
     ["(", ")", bug7610], ["[", "]", bug8133],
     ["{", "}", bug8133], ["<", ">", bug8133],
    ].each do |lparen, rparen, bug|
      s = "\\#{lparen}a\\#{rparen}"
      assert_equal(/#{s}/, eval("%r#{lparen}#{s}#{rparen}"), bug)
    end
  end

  def test_source_unescaped
    expected, result = "!\"#%&',-/:;=@_`~".each_char.map {|c|
      [
        ["#{c}", "\\#{c}", 1],
        begin
          re = eval("%r#{c}\\#{c}#{c}", nil, __FILE__, __LINE__)
          t = eval("%r{\\#{c}}", nil, __FILE__, __LINE__).source
        rescue SyntaxError => e
          [e, t, nil]
        else
          [re.source, t, re =~ "a#{c}a"]
        end
      ]
    }.transpose
    assert_equal(expected, result)
  end

  def test_inspect
    assert_equal('//', //.inspect)
    assert_equal('//i', //i.inspect)
    assert_equal('/\//i', /\//i.inspect)
    assert_equal('/\//i', %r"#{'/'}"i.inspect)
    assert_equal('/\/x/i', /\/x/i.inspect)
    assert_equal('/\x00/i', /#{"\0"}/i.inspect)
    assert_equal("/\n/i", /#{"\n"}/i.inspect)
    s = [0xf1, 0xf2, 0xf3].pack("C*")
    assert_equal('/\/\xF1\xF2\xF3/i', /\/#{s}/i.inspect)
  end

  def test_char_to_option
    assert_equal("BAR", "FOOBARBAZ"[/b../i])
    assert_equal("bar", "foobarbaz"[/  b  .  .  /x])
    assert_equal("bar\n", "foo\nbar\nbaz"[/b.../m])
    assert_raise(SyntaxError) { eval('//z') }
  end

  def test_char_to_option_kcode
    assert_equal("bar", "foobarbaz"[/b../s])
    assert_equal("bar", "foobarbaz"[/b../e])
    assert_equal("bar", "foobarbaz"[/b../u])
  end

  def test_to_s2
    assert_equal('(?-mix:foo)', /(?:foo)/.to_s)
    assert_equal('(?m-ix:foo)', /(?:foo)/m.to_s)
    assert_equal('(?mi-x:foo)', /(?:foo)/mi.to_s)
    assert_equal('(?mix:foo)', /(?:foo)/mix.to_s)
    assert_equal('(?m-ix:foo)', /(?m-ix:foo)/.to_s)
    assert_equal('(?mi-x:foo)', /(?mi-x:foo)/.to_s)
    assert_equal('(?mix:foo)', /(?mix:foo)/.to_s)
    assert_equal('(?mix:)', /(?mix)/.to_s)
    assert_equal('(?-mix:(?mix:foo) )', /(?mix:foo) /.to_s)
  end

  def test_casefold_p
    assert_equal(false, /a/.casefold?)
    assert_equal(true, /a/i.casefold?)
    assert_equal(false, /(?i:a)/.casefold?)
  end

  def test_options
    assert_equal(Regexp::IGNORECASE, /a/i.options)
    assert_equal(Regexp::EXTENDED, /a/x.options)
    assert_equal(Regexp::MULTILINE, /a/m.options)
  end

  def test_match_init_copy
    m = /foo/.match("foo")
    assert_equal(/foo/, m.dup.regexp)
    assert_raise(TypeError) do
      m.instance_eval { initialize_copy(nil) }
    end
    assert_equal([0, 3], m.offset(0))
    assert_equal(/foo/, m.dup.regexp)
  end

  def test_match_size
    m = /(.)(.)(\d+)(\d)/.match("THX1138.")
    assert_equal(5, m.size)
  end

  def test_match_offset_begin_end
    m = /(?<x>b..)/.match("foobarbaz")
    assert_equal([3, 6], m.offset("x"))
    assert_equal(3, m.begin("x"))
    assert_equal(6, m.end("x"))
    assert_raise(IndexError) { m.offset("y") }
    assert_raise(IndexError) { m.offset(2) }
    assert_raise(IndexError) { m.begin(2) }
    assert_raise(IndexError) { m.end(2) }

    m = /(?<x>q..)?/.match("foobarbaz")
    assert_equal([nil, nil], m.offset("x"))
    assert_equal(nil, m.begin("x"))
    assert_equal(nil, m.end("x"))

    m = /\A\u3042(.)(.)?(.)\z/.match("\u3042\u3043\u3044")
    assert_equal([1, 2], m.offset(1))
    assert_equal([nil, nil], m.offset(2))
    assert_equal([2, 3], m.offset(3))
  end

  def test_match_byteoffset_begin_end
    m = /(?<x>b..)/.match("foobarbaz")
    assert_equal([3, 6], m.byteoffset("x"))
    assert_equal(3, m.begin("x"))
    assert_equal(6, m.end("x"))
    assert_raise(IndexError) { m.byteoffset("y") }
    assert_raise(IndexError) { m.byteoffset(2) }
    assert_raise(IndexError) { m.begin(2) }
    assert_raise(IndexError) { m.end(2) }

    m = /(?<x>q..)?/.match("foobarbaz")
    assert_equal([nil, nil], m.byteoffset("x"))
    assert_equal(nil, m.begin("x"))
    assert_equal(nil, m.end("x"))

    m = /\A\u3042(.)(.)?(.)\z/.match("\u3042\u3043\u3044")
    assert_equal([3, 6], m.byteoffset(1))
    assert_equal([nil, nil], m.byteoffset(2))
    assert_equal([6, 9], m.byteoffset(3))
  end

  def test_match_to_s
    m = /(?<x>b..)/.match("foobarbaz")
    assert_equal("bar", m.to_s)
  end

  def test_match_pre_post
    m = /(?<x>b..)/.match("foobarbaz")
    assert_equal("foo", m.pre_match)
    assert_equal("baz", m.post_match)
  end

  def test_match_array
    m = /(...)(...)(...)(...)?/.match("foobarbaz")
    assert_equal(["foobarbaz", "foo", "bar", "baz", nil], m.to_a)
  end

  def test_match_captures
    m = /(...)(...)(...)(...)?/.match("foobarbaz")
    assert_equal(["foo", "bar", "baz", nil], m.captures)
  end

  def test_match_aref
    m = /(...)(...)(...)(...)?/.match("foobarbaz")
    assert_equal("foobarbaz", m[0])
    assert_equal("foo", m[1])
    assert_equal("foo", m[-4])
    assert_nil(m[-1])
    assert_nil(m[-11])
    assert_nil(m[-11, 1])
    assert_nil(m[-11..1])
    assert_nil(m[5])
    assert_nil(m[9])
    assert_equal(["foo", "bar", "baz"], m[1..3])
    assert_equal(["foo", "bar", "baz"], m[1, 3])
    assert_equal([], m[3..1])
    assert_equal([], m[3, 0])
    assert_equal(nil, m[3, -1])
    assert_equal(nil, m[9, 1])
    assert_equal(["baz"], m[3, 1])
    assert_equal(["baz", nil], m[3, 5])
    assert_nil(m[5])
    assert_raise(IndexError) { m[:foo] }
    assert_raise(TypeError) { m[nil] }
    assert_equal(["baz", nil], m[-2, 3])
  end

  def test_match_values_at
    idx = Object.new
    def idx.to_int; 2; end
    m = /(...)(...)(...)(...)?/.match("foobarbaz")
    assert_equal(["foo", "bar", "baz"], m.values_at(1, 2, 3))
    assert_equal(["foo", "bar", "baz"], m.values_at(1..3))
    assert_equal(["foo", "bar", "baz", nil, nil], m.values_at(1..5))
    assert_equal([], m.values_at(3..1))
    assert_equal([nil, nil, nil, nil, nil], m.values_at(5..9))
    assert_equal(["bar"], m.values_at(idx))
    assert_raise(RangeError){ m.values_at(-11..1) }
    assert_raise(TypeError){ m.values_at(nil) }

    m = /(?<a>\d+) *(?<op>[+\-*\/]) *(?<b>\d+)/.match("1 + 2")
    assert_equal(["1", "2", "+"], m.values_at(:a, 'b', :op))
    assert_equal(["+"], m.values_at(idx))
    assert_raise(TypeError){ m.values_at(nil) }
    assert_raise(IndexError){ m.values_at(:foo) }
  end

  def test_match_string
    m = /(?<x>b..)/.match("foobarbaz")
    assert_equal("foobarbaz", m.string)
  end

  def test_match_matchsubstring
    m = /(.)(.)(\d+)(\d)(\w)?/.match("THX1138.")
    assert_equal("HX1138", m.match(0))
    assert_equal("8", m.match(4))
    assert_nil(m.match(5))

    m = /\A\u3042(.)(.)?(.)\z/.match("\u3042\u3043\u3044")
    assert_equal("\u3043", m.match(1))
    assert_nil(m.match(2))
    assert_equal("\u3044", m.match(3))

    m = /(?<foo>.)(?<n>[^aeiou])?(?<bar>.+)/.match("hoge\u3042")
    assert_equal("h", m.match(:foo))
    assert_nil(m.match(:n))
    assert_equal("oge\u3042", m.match(:bar))
  end

  def test_match_match_length
    m = /(.)(.)(\d+)(\d)(\w)?/.match("THX1138.")
    assert_equal(6, m.match_length(0))
    assert_equal(1, m.match_length(4))
    assert_nil(m.match_length(5))

    m = /\A\u3042(.)(.)?(.)\z/.match("\u3042\u3043\u3044")
    assert_equal(1, m.match_length(1))
    assert_nil(m.match_length(2))
    assert_equal(1, m.match_length(3))

    m = /(?<foo>.)(?<n>[^aeiou])?(?<bar>.+)/.match("hoge\u3042")
    assert_equal(1, m.match_length(:foo))
    assert_nil(m.match_length(:n))
    assert_equal(4, m.match_length(:bar))
  end

  def test_match_inspect
    m = /(...)(...)(...)(...)?/.match("foobarbaz")
    assert_equal('#<MatchData "foobarbaz" 1:"foo" 2:"bar" 3:"baz" 4:nil>', m.inspect)
  end

  def test_match_data_deconstruct
    m = /foo.+/.match("foobarbaz")
    assert_equal([], m.deconstruct)

    m = /(foo).+(baz)/.match("foobarbaz")
    assert_equal(["foo", "baz"], m.deconstruct)

    m = /(...)(...)(...)(...)?/.match("foobarbaz")
    assert_equal(["foo", "bar", "baz", nil], m.deconstruct)
  end

  def test_match_data_deconstruct_keys
    m = /foo.+/.match("foobarbaz")
    assert_equal({}, m.deconstruct_keys([:a]))

    m = /(?<a>foo).+(?<b>baz)/.match("foobarbaz")
    assert_equal({a: "foo", b: "baz"}, m.deconstruct_keys(nil))
    assert_equal({a: "foo", b: "baz"}, m.deconstruct_keys([:a, :b]))
    assert_equal({b: "baz"}, m.deconstruct_keys([:b]))
    assert_equal({}, m.deconstruct_keys([:c, :a]))
    assert_equal({a: "foo"}, m.deconstruct_keys([:a, :c]))
    assert_equal({}, m.deconstruct_keys([:a, :b, :c]))

    assert_raise(TypeError) {
      m.deconstruct_keys(0)
    }

    assert_raise(TypeError) {
      m.deconstruct_keys(["a", "b"])
    }
  end

  def test_initialize
    assert_raise(ArgumentError) { Regexp.new }
    assert_equal(/foo/, assert_warning(/ignored/) {Regexp.new(/foo/, Regexp::IGNORECASE)})
    assert_equal(/foo/, assert_no_warning(/ignored/) {Regexp.new(/foo/)})
    assert_equal(/foo/, assert_no_warning(/ignored/) {Regexp.new(/foo/, timeout: nil)})

    arg_encoding_none = //n.options # ARG_ENCODING_NONE is implementation defined value

    assert_deprecated_warning('') do
      assert_equal(Encoding.find("US-ASCII"), Regexp.new("b..", Regexp::NOENCODING).encoding)
      assert_equal("bar", "foobarbaz"[Regexp.new("b..", Regexp::NOENCODING)])
      assert_equal(//, Regexp.new(""))
      assert_equal(//, Regexp.new("", timeout: 1))
      assert_equal(//n, Regexp.new("", Regexp::NOENCODING))
      assert_equal(//n, Regexp.new("", Regexp::NOENCODING, timeout: 1))

      assert_equal(arg_encoding_none, Regexp.new("", Regexp::NOENCODING).options)

      assert_nil(Regexp.new("").timeout)
      assert_equal(1.0, Regexp.new("", timeout: 1.0).timeout)
      assert_nil(Regexp.compile("").timeout)
      assert_equal(1.0, Regexp.compile("", timeout: 1.0).timeout)
    end

    assert_raise(RegexpError) { Regexp.new(")(") }
    assert_raise(RegexpError) { Regexp.new('[\\40000000000') }
    assert_raise(RegexpError) { Regexp.new('[\\600000000000.') }
    assert_raise(RegexpError) { Regexp.new("((?<v>))\\g<0>") }
  end

  def test_initialize_from_regex_memory_corruption
    assert_ruby_status([], <<-'end;')
      10_000.times { Regexp.new(Regexp.new("(?<name>)")) }
    end;
  end

  def test_initialize_bool_warning
    assert_warning(/expected true or false as ignorecase/) do
      Regexp.new("foo", :i)
    end
  end

  def test_initialize_option
    assert_equal(//i, Regexp.new("", "i"))
    assert_equal(//m, Regexp.new("", "m"))
    assert_equal(//x, Regexp.new("", "x"))
    assert_equal(//imx, Regexp.new("", "imx"))
    assert_equal(//, Regexp.new("", ""))
    assert_equal(//imx, Regexp.new("", "mimix"))

    assert_raise(ArgumentError) { Regexp.new("", "e") }
    assert_raise(ArgumentError) { Regexp.new("", "n") }
    assert_raise(ArgumentError) { Regexp.new("", "s") }
    assert_raise(ArgumentError) { Regexp.new("", "u") }
    assert_raise(ArgumentError) { Regexp.new("", "o") }
    assert_raise(ArgumentError) { Regexp.new("", "j") }
    assert_raise(ArgumentError) { Regexp.new("", "xmen") }
  end

  def test_match_control_meta_escape
    assert_equal(0, /\c\xFF/ =~ "\c\xFF")
    assert_equal(0, /\c\M-\xFF/ =~ "\c\M-\xFF")
    assert_equal(0, /\C-\xFF/ =~ "\C-\xFF")
    assert_equal(0, /\C-\M-\xFF/ =~ "\C-\M-\xFF")
    assert_equal(0, /\M-\xFF/ =~ "\M-\xFF")
    assert_equal(0, /\M-\C-\xFF/ =~ "\M-\C-\xFF")
    assert_equal(0, /\M-\c\xFF/ =~ "\M-\c\xFF")

    assert_nil(/\c\xFE/ =~ "\c\xFF")
    assert_nil(/\c\M-\xFE/ =~ "\c\M-\xFF")
    assert_nil(/\C-\xFE/ =~ "\C-\xFF")
    assert_nil(/\C-\M-\xFE/ =~ "\C-\M-\xFF")
    assert_nil(/\M-\xFE/ =~ "\M-\xFF")
    assert_nil(/\M-\C-\xFE/ =~ "\M-\C-\xFF")
    assert_nil(/\M-\c\xFE/ =~ "\M-\c\xFF")
  end

  def test_unescape
    assert_raise(ArgumentError) { s = '\\'; /#{ s }/ }
    assert_equal(/\xFF/n, /#{ s="\\xFF" }/n)
    assert_equal(/\177/, (s = '\177'; /#{ s }/))
    assert_raise(ArgumentError) { s = '\u'; /#{ s }/ }
    assert_raise(ArgumentError) { s = '\u{ ffffffff }'; /#{ s }/ }
    assert_raise(ArgumentError) { s = '\u{ ffffff }'; /#{ s }/ }
    assert_raise(ArgumentError) { s = '\u{ ffff X }'; /#{ s }/ }
    assert_raise(ArgumentError) { s = '\u{ }'; /#{ s }/ }
    assert_equal("b", "abc"[(s = '\u{0062}'; /#{ s }/)])
    assert_equal("b", "abc"[(s = '\u0062'; /#{ s }/)])
    assert_raise(ArgumentError) { s = '\u0'; /#{ s }/ }
    assert_raise(ArgumentError) { s = '\u000X'; /#{ s }/ }
    assert_raise(ArgumentError) { s = "\xff" + '\u3042'; /#{ s }/ }
    assert_raise(ArgumentError) { s = '\u3042' + [0xff].pack("C"); /#{ s }/ }
    assert_raise(SyntaxError) { s = ''; eval(%q(/\u#{ s }/)) }

    assert_equal(/a/, eval(%q(s="\u0061";/#{s}/n)))
    assert_raise(RegexpError) { s = "\u3042"; eval(%q(/#{s}/n)) }
    assert_raise(RegexpError) { s = "\u0061"; eval(%q(/\u3042#{s}/n)) }
    assert_raise(RegexpError) { s1=[0xff].pack("C"); s2="\u3042"; eval(%q(/#{s1}#{s2}/)); [s1, s2] }

    assert_raise(ArgumentError) { s = '\x'; /#{ s }/ }

    assert_equal("\xe1", [0x00, 0xe1, 0xff].pack("C*")[/\M-a/])
    assert_equal("\xdc", [0x00, 0xdc, 0xff].pack("C*")[/\M-\\/])
    assert_equal("\x8a", [0x00, 0x8a, 0xff].pack("C*")[/\M-\n/])
    assert_equal("\x89", [0x00, 0x89, 0xff].pack("C*")[/\M-\t/])
    assert_equal("\x8d", [0x00, 0x8d, 0xff].pack("C*")[/\M-\r/])
    assert_equal("\x8c", [0x00, 0x8c, 0xff].pack("C*")[/\M-\f/])
    assert_equal("\x8b", [0x00, 0x8b, 0xff].pack("C*")[/\M-\v/])
    assert_equal("\x87", [0x00, 0x87, 0xff].pack("C*")[/\M-\a/])
    assert_equal("\x9b", [0x00, 0x9b, 0xff].pack("C*")[/\M-\e/])
    assert_equal("\x01", [0x00, 0x01, 0xff].pack("C*")[/\C-a/])

    assert_raise(ArgumentError) { s = '\M'; /#{ s }/ }
    assert_raise(ArgumentError) { s = '\M-\M-a'; /#{ s }/ }
    assert_raise(ArgumentError) { s = '\M-\\'; /#{ s }/ }

    assert_raise(ArgumentError) { s = '\C'; /#{ s }/ }
    assert_raise(ArgumentError) { s = '\c'; /#{ s }/ }
    assert_raise(ArgumentError) { s = '\C-\C-a'; /#{ s }/ }

    assert_raise(ArgumentError) { s = '\M-\z'; /#{ s }/ }
    assert_raise(ArgumentError) { s = '\M-\777'; /#{ s }/ }

    assert_equal("\u3042\u3042", "\u3042\u3042"[(s = "\u3042" + %q(\xe3\x81\x82); /#{s}/)])
    assert_raise(ArgumentError) { s = "\u3042" + %q(\xe3); /#{s}/ }
    assert_raise(ArgumentError) { s = "\u3042" + %q(\xe3\xe3); /#{s}/ }
    assert_raise(ArgumentError) { s = '\u3042' + [0xff].pack("C"); /#{s}/ }

    assert_raise(SyntaxError) { eval("/\u3042/n") }

    s = ".........."
    5.times { s.sub!(".", "") }
    assert_equal(".....", s)

    assert_equal("\\\u{3042}", Regexp.new("\\\u{3042}").source)
  end

  def test_equal
    bug5484 = '[ruby-core:40364]'
    assert_equal(/abc/, /abc/)
    assert_not_equal(/abc/, /abc/m)
    assert_not_equal(/abc/, /abd/)
    assert_equal(/\/foo/, Regexp.new('/foo'), bug5484)
  end

  def test_match
    assert_nil(//.match(nil))
    assert_equal("abc", /.../.match(:abc)[0])
    assert_raise(TypeError) { /.../.match(Object.new)[0] }
    assert_equal("bc", /../.match('abc', 1)[0])
    assert_equal("bc", /../.match('abc', -2)[0])
    assert_nil(/../.match("abc", -4))
    assert_nil(/../.match("abc", 4))

    # use eval because only one warning is shown for the same regexp literal
    pat = eval('/../n')
    assert_equal('\x', assert_warning(/binary regexp/) {pat.match("\u3042" + '\x', 1)}[0])

    r = nil
    /.../.match("abc") {|m| r = m[0] }
    assert_equal("abc", r)

    $_ = "abc"; assert_equal(1, ~/bc/)
    $_ = "abc"; assert_nil(~/d/)
    $_ = nil; assert_nil(~/./)
  end

  def test_match_p
    /backref/ =~ 'backref'
    # must match here, but not in a separate method, e.g., assert_send,
    # to check if $~ is affected or not.
    assert_equal(false, //.match?(nil))
    assert_equal(true, //.match?(""))
    assert_equal(true, /.../.match?(:abc))
    assert_raise(TypeError) { /.../.match?(Object.new) }
    assert_equal(true, /b/.match?('abc'))
    assert_equal(true, /b/.match?('abc', 1))
    assert_equal(true, /../.match?('abc', 1))
    assert_equal(true, /../.match?('abc', -2))
    assert_equal(false, /../.match?("abc", -4))
    assert_equal(false, /../.match?("abc", 4))
    assert_equal(true, /../.match?("\u3042xx", 1))
    assert_equal(false, /../.match?("\u3042x", 1))
    assert_equal(true, /\z/.match?(""))
    assert_equal(true, /\z/.match?("abc"))
    assert_equal(true, /R.../.match?("Ruby"))
    assert_equal(false, /R.../.match?("Ruby", 1))
    assert_equal(false, /P.../.match?("Ruby"))
    assert_equal('backref', $&)
  end

  def test_eqq
    assert_equal(false, /../ === nil)
  end

  def test_quote
    assert_equal("\xff", Regexp.quote([0xff].pack("C")))
    assert_equal("\\ ", Regexp.quote("\ "))
    assert_equal("\\t", Regexp.quote("\t"))
    assert_equal("\\n", Regexp.quote("\n"))
    assert_equal("\\r", Regexp.quote("\r"))
    assert_equal("\\f", Regexp.quote("\f"))
    assert_equal("\\v", Regexp.quote("\v"))
    assert_equal("\u3042\\t", Regexp.quote("\u3042\t"))
    assert_equal("\\t\xff", Regexp.quote("\t" + [0xff].pack("C")))

    bug13034 = '[ruby-core:78646] [Bug #13034]'
    str = "\x00".force_encoding("UTF-16BE")
    assert_equal(str, Regexp.quote(str), bug13034)
  end

  def test_try_convert
    assert_equal(/re/, Regexp.try_convert(/re/))
    assert_nil(Regexp.try_convert("re"))

    o = Object.new
    assert_nil(Regexp.try_convert(o))
    def o.to_regexp() /foo/ end
    assert_equal(/foo/, Regexp.try_convert(o))
  end

  def test_union2
    assert_equal(/(?!)/, Regexp.union)
    assert_equal(/foo/, Regexp.union(/foo/))
    assert_equal(/foo/, Regexp.union([/foo/]))
    assert_equal(/\t/, Regexp.union("\t"))
    assert_equal(/(?-mix:\u3042)|(?-mix:\u3042)/, Regexp.union(/\u3042/, /\u3042/))
    assert_equal("\u3041", "\u3041"[Regexp.union(/\u3042/, "\u3041")])
  end

  def test_dup
    assert_equal(//, //.dup)
    assert_raise(TypeError) { //.dup.instance_eval { initialize_copy(nil) } }
  end

  def test_regsub
    assert_equal("fooXXXbaz", "foobarbaz".sub!(/bar/, "XXX"))
    s = [0xff].pack("C")
    assert_equal(s, "X".sub!(/./, s))
    assert_equal('\\' + s, "X".sub!(/./, '\\' + s))
    assert_equal('\k', "foo".sub!(/.../, '\k'))
    assert_raise(RuntimeError) { "foo".sub!(/(?<x>o)/, '\k<x') }
    assert_equal('foo[bar]baz', "foobarbaz".sub!(/(b..)/, '[\0]'))
    assert_equal('foo[foo]baz', "foobarbaz".sub!(/(b..)/, '[\`]'))
    assert_equal('foo[baz]baz', "foobarbaz".sub!(/(b..)/, '[\\\']'))
    assert_equal('foo[r]baz', "foobarbaz".sub!(/(b)(.)(.)/, '[\+]'))
    assert_equal('foo[\\]baz', "foobarbaz".sub!(/(b..)/, '[\\\\]'))
    assert_equal('foo[\z]baz', "foobarbaz".sub!(/(b..)/, '[\z]'))
  end

  def test_regsub_K
    bug8856 = '[ruby-dev:47694] [Bug #8856]'
    result = "foobarbazquux/foobarbazquux".gsub(/foo\Kbar/, "")
    assert_equal('foobazquux/foobazquux', result, bug8856)
  end

  def test_ignorecase
    v = assert_deprecated_warning(/variable \$= is no longer effective/) { $= }
    assert_equal(false, v)
    assert_deprecated_warning(/variable \$= is no longer effective; ignored/) { $= = nil }
  end

  def test_match_setter
    /foo/ =~ "foo"
    m = $~
    /bar/ =~ "bar"
    $~ = m
    assert_equal("foo", $&)
  end

  def test_match_without_regexp
    # create a MatchData for each assertion because the internal state may change
    test = proc {|&blk| "abc".sub("a", ""); blk.call($~) }

    bug10877 = '[ruby-core:68209] [Bug #10877]'
    bug18160 = '[Bug #18160]'
    test.call {|m| assert_raise_with_message(IndexError, /foo/, bug10877) {m["foo"]} }
    key = "\u{3042}"
    [Encoding::UTF_8, Encoding::Shift_JIS, Encoding::EUC_JP].each do |enc|
      idx = key.encode(enc)
      pat = /#{idx}/
      test.call {|m| assert_raise_with_message(IndexError, pat, bug10877) {m[idx]} }
      test.call {|m| assert_raise_with_message(IndexError, pat, bug18160) {m.offset(idx)} }
      test.call {|m| assert_raise_with_message(IndexError, pat, bug18160) {m.begin(idx)} }
      test.call {|m| assert_raise_with_message(IndexError, pat, bug18160) {m.end(idx)} }
    end
    test.call {|m| assert_equal(/a/, m.regexp) }
    test.call {|m| assert_equal("abc", m.string) }
    test.call {|m| assert_equal(1, m.size) }
    test.call {|m| assert_equal(0, m.begin(0)) }
    test.call {|m| assert_equal(1, m.end(0)) }
    test.call {|m| assert_equal([0, 1], m.offset(0)) }
    test.call {|m| assert_equal([], m.captures) }
    test.call {|m| assert_equal([], m.names) }
    test.call {|m| assert_equal({}, m.named_captures) }
    test.call {|m| assert_equal(/a/.match("abc"), m) }
    test.call {|m| assert_equal(/a/.match("abc").hash, m.hash) }
    test.call {|m| assert_equal("bc", m.post_match) }
    test.call {|m| assert_equal("", m.pre_match) }
    test.call {|m| assert_equal(["a", nil], m.values_at(0, 1)) }
  end

  def test_last_match
    /(...)(...)(...)(...)?/.match("foobarbaz")
    assert_equal("foobarbaz", Regexp.last_match(0))
    assert_equal("foo", Regexp.last_match(1))
    assert_nil(Regexp.last_match(5))
    assert_nil(Regexp.last_match(-1))
  end

  def test_getter
    alias $__REGEXP_TEST_LASTMATCH__ $&
    alias $__REGEXP_TEST_PREMATCH__ $`
    alias $__REGEXP_TEST_POSTMATCH__ $'
    alias $__REGEXP_TEST_LASTPARENMATCH__ $+
    /(b)(.)(.)/.match("foobarbaz")
    assert_equal("bar", $__REGEXP_TEST_LASTMATCH__)
    assert_equal("foo", $__REGEXP_TEST_PREMATCH__)
    assert_equal("baz", $__REGEXP_TEST_POSTMATCH__)
    assert_equal("r", $__REGEXP_TEST_LASTPARENMATCH__)

    /(...)(...)(...)/.match("foobarbaz")
    assert_equal("baz", $+)
  end

  def test_rindex_regexp
    # use eval because only one warning is shown for the same regexp literal
    pat = eval('/b../n')
    assert_equal(3, assert_warning(/binary regexp/) {"foobarbaz\u3042".rindex(pat, 5)})
  end

  def assert_regexp(re, ss, fs = [], msg = nil)
    re = EnvUtil.suppress_warning {Regexp.new(re)} unless re.is_a?(Regexp)
    ss = [ss] unless ss.is_a?(Array)
    ss.each do |e, s|
      s ||= e
      assert_match(re, s, msg)
      m = re.match(s)
      assert_equal(e, m[0], msg)
    end
    fs = [fs] unless fs.is_a?(Array)
    fs.each {|s| assert_no_match(re, s, msg) }
  end
  alias check assert_regexp

  def assert_fail(re)
    assert_raise(RegexpError) { %r"#{ re }" }
  end
  alias failcheck assert_fail

  def test_parse
    check(/\*\+\?\{\}\|\(\)\<\>\`\'/, "*+?{}|()<>`'")
    check(/\A\w\W\z/, %w(a. b!), %w(.. ab))
    check(/\A.\b.\b.\B.\B.\z/, %w(a.aaa .a...), %w(aaaaa .....))
    check(/\A\s\S\z/, [' a', "\n."], ['  ', "\n\n", 'a '])
    check(/\A\d\D\z/, '0a', %w(00 aa))
    check(/\A\h\H\z/, %w(0g ag BH), %w(a0 af GG))
    check(/\Afoo\Z\s\z/, "foo\n", ["foo", "foo\nbar"])
    assert_equal(%w(a b c), "abc def".scan(/\G\w/))
    check(/\A\u3042\z/, "\u3042", ["", "\u3043", "a"])
    check(/\A(..)\1\z/, %w(abab ....), %w(abba aba))
    failcheck('\1')
    check(/\A\80\z/, "80", ["\100", ""])
    check(/\A\77\z/, "?")
    check(/\A\78\z/, "\7" + '8', ["\100", ""])
    check(assert_warning(/Unknown escape/) {eval('/\A\Qfoo\E\z/')}, "QfooE")
    check(/\Aa++\z/, "aaa")
    check('\Ax]\z', "x]")
    check(/x#foo/x, "x", "#foo")
    check(/\Ax#foo#{ "\n" }x\z/x, "xx", ["x", "x#foo\nx"])
    check(/\A\p{Alpha}\z/, ["a", "z"], [".", "", ".."])
    check(/\A\p{^Alpha}\z/, [".", "!"], ["!a", ""])
    check(/\A\n\z/, "\n")
    check(/\A\t\z/, "\t")
    check(/\A\r\z/, "\r")
    check(/\A\f\z/, "\f")
    check(/\A\a\z/, "\007")
    check(/\A\e\z/, "\033")
    check(/\A\v\z/, "\v")
    failcheck('(')
    failcheck('(?foo)')
    failcheck('/\p{foobarbazqux}/')
    failcheck('/\p{foobarbazqux' + 'a' * 1000 + '}/')
    failcheck('/[1-\w]/')
  end

  def test_exec
    check(/A*B/, %w(B AB AAB AAAB), %w(A))
    check(/\w*!/, %w(! a! ab! abc!), %w(abc))
    check(/\w*\W/, %w(! a" ab# abc$), %w(abc))
    check(/\w*\w/, %w(z az abz abcz), %w(!))
    check(/[a-z]*\w/, %w(z az abz abcz), %w(!))
    check(/[a-z]*\W/, %w(! a" ab# abc$), %w(A))
    check(/((a|bb|ccc|dddd)(1|22|333|4444))/i, %w(a1 bb1 a22), %w(a2 b1))
    check(/\u0080/, (1..4).map {|i| ["\u0080", "\u0080" * i] }, ["\u0081"])
    check(/\u0080\u0080/, (2..4).map {|i| ["\u0080" * 2, "\u0080" * i] }, ["\u0081"])
    check(/\u0080\u0080\u0080/, (3..4).map {|i| ["\u0080" * 3, "\u0080" * i] }, ["\u0081"])
    check(/\u0080\u0080\u0080\u0080/, (4..4).map {|i| ["\u0080" * 4, "\u0080" * i] }, ["\u0081"])
    check(/[^\u3042\u3043\u3044]/, %W(a b \u0080 \u3041 \u3045), %W(\u3042 \u3043 \u3044))
    check(/a.+/m, %W(a\u0080 a\u0080\u0080 a\u0080\u0080\u0080), %W(a))
    check(/a.+z/m, %W(a\u0080z a\u0080\u0080z a\u0080\u0080\u0080z), %W(az))
    check(/abc\B.\Bxyz/, %w(abcXxyz abc0xyz), %w(abc|xyz abc-xyz))
    check(/\Bxyz/, [%w(xyz abcXxyz), %w(xyz abc0xyz)], %w(abc xyz abc-xyz))
    check(/abc\B/, [%w(abc abcXxyz), %w(abc abc0xyz)], %w(abc xyz abc-xyz))
    failcheck('(?<foo>abc)\1')
    check(/^(A+|B+)(?>\g<1>)*[BC]$/, %w(AC BC ABC BAC AABBC), %w(AABB))
    check(/^(A+|B(?>\g<1>)*)[AC]$/, %w(AAAC BBBAAAAC), %w(BBBAAA))
    check(/^()(?>\g<1>)*$/, "", "a")
    check(/^(?>(?=a)(#{ "a" * 1000 }|))++$/, ["a" * 1000, "a" * 2000, "a" * 3000], ["", "a" * 500, "b" * 1000])
    check(assert_warning(/nested repeat operator/) {eval('/^(?:a?)?$/')}, ["", "a"], ["aa"])
    check(assert_warning(/nested repeat operator/) {eval('/^(?:a+)?$/')}, ["", "a", "aa"], ["ab"])
    check(/^(?:a?)+?$/, ["", "a", "aa"], ["ab"])
    check(/^a??[ab]/, [["a", "a"], ["a", "aa"], ["b", "b"], ["a", "ab"]], ["c"])
    check(/^(?:a*){3,5}$/, ["", "a", "aa", "aaa", "aaaa", "aaaaa", "aaaaaa"], ["b"])
    check(/^(?:a+){3,5}$/, ["aaa", "aaaa", "aaaaa", "aaaaaa"], ["", "a", "aa", "b"])
  end

  def test_parse_look_behind
    check(/(?<=A)B(?=C)/, [%w(B ABC)], %w(aBC ABc aBc))
    check(/(?<!A)B(?!C)/, [%w(B aBc)], %w(ABC aBC ABc))
    failcheck('(?<=.*)')
    failcheck('(?<!.*)')
    check(/(?<=A|B.)C/, [%w(C AC), %w(C BXC)], %w(C BC))
    check(/(?<!A|B.)C/, [%w(C C), %w(C BC)], %w(AC BXC))

    assert_not_match(/(?<!aa|b)c/i, "Aac")
    assert_not_match(/(?<!b|aa)c/i, "Aac")
  end

  def test_parse_kg
    check(/\A(.)(.)\k<1>(.)\z/, %w(abac abab ....), %w(abcd aaba xxx))
    check(/\A(.)(.)\k<-1>(.)\z/, %w(abbc abba ....), %w(abcd aaba xxx))
    check(/\A(?<n>.)(?<x>\g<n>){0}(?<y>\k<n+0>){0}\g<x>\g<y>\z/, "aba", "abb")
    check(/\A(?<n>.)(?<x>\g<n>){0}(?<y>\k<n+1>){0}\g<x>\g<y>\z/, "abb", "aba")
    check(/\A(?<x>..)\k<x>\z/, %w(abab ....), %w(abac abba xxx))
    check(/\A(.)(..)\g<-1>\z/, "abcde", %w(.... ......))
    failcheck('\k<x>')
    failcheck('\k<')
    failcheck('\k<>')
    failcheck('\k<.>')
    failcheck('\k<x.>')
    failcheck('\k<1.>')
    failcheck('\k<x')
    failcheck('\k<x+')
    failcheck('()\k<-2>')
    failcheck('()\g<-2>')
    check(/\A(?<x>.)(?<x>.)\k<x>\z/, %w(aba abb), %w(abc .. ....))
    check(/\A(?<x>.)(?<x>.)\k<x>\z/i, %w(aba ABa abb ABb), %w(abc .. ....))
    check('\k\g', "kg")
    failcheck('(.\g<1>)')
    failcheck('(.\g<2>)')
    failcheck('(?=\g<1>)')
    failcheck('((?=\g<1>))')
    failcheck('(\g<1>|.)')
    failcheck('(.|\g<1>)')
    check(/(!)(?<=(a)|\g<1>)/, ["!"], %w(a))
    check(/^(a|b\g<1>c)$/, %w(a bac bbacc bbbaccc), %w(bbac bacc))
    check(/^(a|b\g<2>c)(B\g<1>C){0}$/, %w(a bBaCc bBbBaCcCc bBbBbBaCcCcCc), %w(bBbBaCcC BbBaCcCc))
    check(/\A(?<n>.|X\g<n>)(?<x>\g<n>){0}(?<y>\k<n+0>){0}\g<x>\g<y>\z/, "XXaXbXXa", %w(XXabXa abb))
    check(/\A(?<n>.|X\g<n>)(?<x>\g<n>){0}(?<y>\k<n+1>){0}\g<x>\g<y>\z/, "XaXXbXXb", %w(aXXbXb aba))
    failcheck('(?<x>)(?<x>)(\g<x>)')
    check(/^(?<x>foo)(bar)\k<x>/, %w(foobarfoo), %w(foobar barfoo))
    check(/^(?<a>f)(?<a>o)(?<a>o)(?<a>b)(?<a>a)(?<a>r)(?<a>b)(?<a>a)(?<a>z)\k<a>{9}$/, %w(foobarbazfoobarbaz foobarbazbazbarfoo foobarbazzabraboof), %w(foobar barfoo))
  end

  def test_parse_curly_brace
    check(/\A{/, ["{", ["{", "{x"]])
    check(/\A{ /, ["{ ", ["{ ", "{ x"]])
    check(/\A{,}\z/, "{,}")
    check(/\A{}\z/, "{}")
    check(/\Aa{0}+\z/, "", %w(a aa aab))
    check(/\Aa{1}+\z/, %w(a aa), ["", "aab"])
    check(/\Aa{1,2}b{1,2}\z/, %w(ab aab abb aabb), ["", "aaabb", "abbb"])
    check(/(?!x){0,1}/, [ ['', 'ab'], ['', ''] ])
    check(/c\z{0,1}/, [ ['c', 'abc'], ['c', 'cab']], ['abd'])
    check(/\A{0,1}a/, [ ['a', 'abc'], ['a', '____abc']], ['bcd'])
    failcheck('.{100001}')
    failcheck('.{0,100001}')
    failcheck('.{1,0}')
    failcheck('{0}')
  end

  def test_parse_comment
    check(/\A(?#foo\)bar)\z/, "", "a")
    failcheck('(?#')
  end

  def test_char_type
    check(/\u3042\d/, ["\u30421", "\u30422"])

    # CClassTable cache test
    assert_match(/\u3042\d/, "\u30421")
    assert_match(/\u3042\d/, "\u30422")
  end

  def test_char_class
    failcheck('[]')
    failcheck('[x')
    check('\A[]]\z', "]", "")
    check('\A[]\.]+\z', %w(] . ]..]), ["", "["])
    check(/\A[\u3042]\z/, "\u3042", "\u3042aa")
    check(/\A[\u3042\x61]+\z/, ["aa\u3042aa", "\u3042\u3042", "a"], ["", "b"])
    check(/\A[\u3042\x61\x62]+\z/, "abab\u3042abab\u3042")
    check(/\A[abc]+\z/, "abcba", ["", "ada"])
    check(/\A[\w][\W]\z/, %w(a. b!), %w(.. ab))
    check(/\A[\s][\S]\z/, [' a', "\n."], ['  ', "\n\n", 'a '])
    check(/\A[\d][\D]\z/, '0a', %w(00 aa))
    check(/\A[\h][\H]\z/, %w(0g ag BH), %w(a0 af GG))
    check(/\A[\p{Alpha}]\z/, ["a", "z"], [".", "", ".."])
    check(/\A[\p{^Alpha}]\z/, [".", "!"], ["!a", ""])
    check(/\A[\xff]\z/, "\xff", ["", "\xfe"])
    check(/\A[\80]+\z/, "8008", ["\\80", "\100", "\1000"])
    check(/\A[\77]+\z/, "???")
    check(/\A[\78]+\z/, "\788\7")
    check(/\A[\0]\z/, "\0")
    check(/\A[[:0]]\z/, [":", "0"], ["", ":0"])
    check(/\A[0-]\z/, ["0", "-"], "0-")
    check('\A[a-&&\w]\z', "a", "-")
    check('\A[--0]\z', ["-", "/", "0"], ["", "1"])
    check('\A[\'--0]\z', %w(* + \( \) 0 ,), ["", ".", "1"])
    check(/\A[a-b-]\z/, %w(a b -), ["", "c"])
    check('\A[a-b-&&\w]\z', %w(a b), ["", "-"])
    check('\A[a-b-&&\W]\z', "-", ["", "a", "b"])
    check('\A[a-c-e]\z', %w(a b c e -), %w(d))
    check(/\A[a-f&&[^b-c]&&[^e]]\z/, %w(a d f), %w(b c e g 0))
    check(/\A[[^b-c]&&[^e]&&a-f]\z/, %w(a d f), %w(b c e g 0))
    check(/\A[\n\r\t]\z/, ["\n", "\r", "\t"])
    failcheck('[9-1]')

    assert_match(/\A\d+\z/, "0123456789")
    assert_no_match(/\d/, "\uff10\uff11\uff12\uff13\uff14\uff15\uff16\uff17\uff18\uff19")
    assert_match(/\A\w+\z/, "09azAZ_")
    assert_no_match(/\w/, "\uff10\uff19\uff41\uff5a\uff21\uff3a")
    assert_match(/\A\s+\z/, "\r\n\v\f\r\s")
    assert_no_match(/\s/, "\u0085")
  end

  def test_posix_bracket
    check(/\A[[:alpha:]0]\z/, %w(0 a), %w(1 .))
    check(assert_warning(/duplicated range/) {eval('/\A[[:^alpha:]0]\z/')}, %w(0 1 .), "a")
    check(assert_warning(/duplicated range/) {eval('/\A[[:alpha\:]]\z/')}, %w(a l p h a :), %w(b 0 1 .))
    check(assert_warning(/duplicated range/) {eval('/\A[[:alpha:foo]0]\z/')}, %w(0 a), %w(1 .))
    check(/\A[[:xdigit:]&&[:alpha:]]\z/, "a", %w(g 0))
    check('\A[[:abcdefghijklmnopqrstu:]]+\z', "[]")
    failcheck('[[:alpha')
    assert_warning(/duplicated range/) {failcheck('[[:alpha:')}
    failcheck('[[:alp:]]')

    assert_match(/\A[[:digit:]]+\z/, "\uff10\uff11\uff12\uff13\uff14\uff15\uff16\uff17\uff18\uff19")
    assert_match(/\A[[:alnum:]]+\z/, "\uff10\uff19\uff41\uff5a\uff21\uff3a")
    assert_match(/\A[[:space:]]+\z/, "\r\n\v\f\r\s\u0085")
    assert_match(/\A[[:ascii:]]+\z/, "\x00\x7F")
    assert_no_match(/[[:ascii:]]/, "\x80\xFF")
  end

  def test_cclass_R
    assert_match(/\A\R\z/, "\r")
    assert_match(/\A\R\z/, "\n")
    assert_match(/\A\R\z/, "\f")
    assert_match(/\A\R\z/, "\v")
    assert_match(/\A\R\z/, "\r\n")
    assert_match(/\A\R\z/, "\u0085")
    assert_match(/\A\R\z/, "\u2028")
    assert_match(/\A\R\z/, "\u2029")
  end

  def test_cclass_X
    assert_match(/\A\X\z/, "\u{20 200d}")
    assert_match(/\A\X\z/, "\u{600 600}")
    assert_match(/\A\X\z/, "\u{600 20}")
    assert_match(/\A\X\z/, "\u{261d 1F3FB}")
    assert_match(/\A\X\z/, "\u{1f600}")
    assert_match(/\A\X\z/, "\u{20 324}")
    assert_match(/\A\X\X\z/, "\u{a 324}")
    assert_match(/\A\X\X\z/, "\u{d 324}")
    assert_match(/\A\X\z/, "\u{1F477 1F3FF 200D 2640 FE0F}")
    assert_match(/\A\X\z/, "\u{1F468 200D 1F393}")
    assert_match(/\A\X\z/, "\u{1F46F 200D 2642 FE0F}")
    assert_match(/\A\X\z/, "\u{1f469 200d 2764 fe0f 200d 1f469}")

    assert_warning('') {/\X/ =~ "\u{a0}"}
  end

  def test_backward
    assert_equal(3, "foobar".rindex(/b.r/i))
    assert_equal(nil, "foovar".rindex(/b.r/i))
    assert_equal(3, ("foo" + "bar" * 1000).rindex(/#{"bar"*1000}/))
    assert_equal(4, ("foo\nbar\nbaz\n").rindex(/bar/i))
  end

  def test_uninitialized
    assert_raise(TypeError) { Regexp.allocate.hash }
    assert_raise(TypeError) { Regexp.allocate.eql? Regexp.allocate }
    assert_raise(TypeError) { Regexp.allocate == Regexp.allocate }
    assert_raise(TypeError) { Regexp.allocate =~ "" }
    assert_equal(false, Regexp.allocate === Regexp.allocate)
    assert_nil(~Regexp.allocate)
    assert_raise(TypeError) { Regexp.allocate.match("") }
    assert_raise(TypeError) { Regexp.allocate.to_s }
    assert_match(/^#<Regexp:.*>$/, Regexp.allocate.inspect)
    assert_raise(TypeError) { Regexp.allocate.source }
    assert_raise(TypeError) { Regexp.allocate.casefold? }
    assert_raise(TypeError) { Regexp.allocate.options }
    assert_equal(Encoding.find("ASCII-8BIT"), Regexp.allocate.encoding)
    assert_equal(false, Regexp.allocate.fixed_encoding?)
    assert_raise(TypeError) { Regexp.allocate.names }
    assert_raise(TypeError) { Regexp.allocate.named_captures }

    assert_not_respond_to(MatchData, :allocate)
=begin
    assert_raise(TypeError) { MatchData.allocate.hash }
    assert_raise(TypeError) { MatchData.allocate.regexp }
    assert_raise(TypeError) { MatchData.allocate.names }
    assert_raise(TypeError) { MatchData.allocate.size }
    assert_raise(TypeError) { MatchData.allocate.length }
    assert_raise(TypeError) { MatchData.allocate.offset(0) }
    assert_raise(TypeError) { MatchData.allocate.begin(0) }
    assert_raise(TypeError) { MatchData.allocate.end(0) }
    assert_raise(TypeError) { MatchData.allocate.to_a }
    assert_raise(TypeError) { MatchData.allocate[:foo] }
    assert_raise(TypeError) { MatchData.allocate.captures }
    assert_raise(TypeError) { MatchData.allocate.values_at }
    assert_raise(TypeError) { MatchData.allocate.pre_match }
    assert_raise(TypeError) { MatchData.allocate.post_match }
    assert_raise(TypeError) { MatchData.allocate.to_s }
    assert_match(/^#<MatchData:.*>$/, MatchData.allocate.inspect)
    assert_raise(TypeError) { MatchData.allocate.string }
    $~ = MatchData.allocate
    assert_raise(TypeError) { $& }
    assert_raise(TypeError) { $` }
    assert_raise(TypeError) { $' }
    assert_raise(TypeError) { $+ }
=end
  end

  def test_unicode
    assert_match(/^\u3042{0}\p{Any}$/, "a")
    assert_match(/^\u3042{0}\p{Any}$/, "\u3041")
    assert_match(/^\u3042{0}\p{Any}$/, "\0")
    assert_match(/^\p{Lo}{4}$/u, "\u3401\u4E01\u{20001}\u{2A701}")
    assert_no_match(/^\u3042{0}\p{Any}$/, "\0\0")
    assert_no_match(/^\u3042{0}\p{Any}$/, "")
    assert_raise(SyntaxError) { eval('/^\u3042{0}\p{' + "\u3042" + '}$/') }
    assert_raise(SyntaxError) { eval('/^\u3042{0}\p{' + 'a' * 1000 + '}$/') }
    assert_raise(SyntaxError) { eval('/^\u3042{0}\p{foobarbazqux}$/') }
    assert_match(/^(\uff21)(a)\1\2$/i, "\uff21A\uff41a")
    assert_no_match(/^(\uff21)\1$/i, "\uff21A")
    assert_no_match(/^(\uff41)\1$/i, "\uff41a")
    assert_match(/^\u00df$/i, "\u00df")
    assert_match(/^\u00df$/i, "ss")
    #assert_match(/^(\u00df)\1$/i, "\u00dfss") # this must be bug...
    assert_match(/^\u00df{2}$/i, "\u00dfss")
    assert_match(/^\u00c5$/i, "\u00c5")
    assert_match(/^\u00c5$/i, "\u00e5")
    assert_match(/^\u00c5$/i, "\u212b")
    assert_match(/^(\u00c5)\1\1$/i, "\u00c5\u00e5\u212b")
    assert_match(/^\u0149$/i, "\u0149")
    assert_match(/^\u0149$/i, "\u02bcn")
    #assert_match(/^(\u0149)\1$/i, "\u0149\u02bcn") # this must be bug...
    assert_match(/^\u0149{2}$/i, "\u0149\u02bcn")
    assert_match(/^\u0390$/i, "\u0390")
    assert_match(/^\u0390$/i, "\u03b9\u0308\u0301")
    #assert_match(/^(\u0390)\1$/i, "\u0390\u03b9\u0308\u0301") # this must be bug...
    assert_match(/^\u0390{2}$/i, "\u0390\u03b9\u0308\u0301")
    assert_match(/^\ufb05$/i, "\ufb05")
    assert_match(/^\ufb05$/i, "\ufb06")
    assert_match(/^\ufb05$/i, "st")
    #assert_match(/^(\ufb05)\1\1$/i, "\ufb05\ufb06st") # this must be bug...
    assert_match(/^\ufb05{3}$/i, "\ufb05\ufb06st")
    assert_match(/^\u03b9\u0308\u0301$/i, "\u0390")
  end

  def test_unicode_age
    assert_unicode_age("\u261c", matches: %w"6.0 1.1", unmatches: [])

    assert_unicode_age("\u31f6", matches: %w"6.0 3.2", unmatches: %w"3.1 3.0 1.1")
    assert_unicode_age("\u2754", matches: %w"6.0", unmatches: %w"5.0 4.0 3.0 2.0 1.1")

    assert_unicode_age("\u32FF", matches: %w"12.1", unmatches: %w"12.0")
  end

  def test_unicode_age_14_0
    @matches = %w"14.0"
    @unmatches = %w"13.0"

    assert_unicode_age("\u{10570}")
    assert_unicode_age("\u9FFF")
    assert_unicode_age("\u{2A6DF}")
    assert_unicode_age("\u{2B738}")
  end

  def test_unicode_age_15_0
    @matches = %w"15.0"
    @unmatches = %w"14.0"

    assert_unicode_age("\u{0CF3}",
                       "KANNADA SIGN COMBINING ANUSVARA ABOVE RIGHT")
    assert_unicode_age("\u{0ECE}", "LAO YAMAKKAN")
    assert_unicode_age("\u{10EFD}".."\u{10EFF}",
                       "ARABIC SMALL LOW WORD SAKTA..ARABIC SMALL LOW WORD MADDA")
    assert_unicode_age("\u{1123F}".."\u{11241}",
                       "KHOJKI LETTER QA..KHOJKI VOWEL SIGN VOCALIC R")
    assert_unicode_age("\u{11B00}".."\u{11B09}",
                       "DEVANAGARI HEAD MARK..DEVANAGARI SIGN MINDU")
    assert_unicode_age("\u{11F00}".."\u{11F10}",
                       "KAWI SIGN CANDRABINDU..KAWI LETTER O")
    assert_unicode_age("\u{11F12}".."\u{11F3A}",
                       "KAWI LETTER KA..KAWI VOWEL SIGN VOCALIC R")
    assert_unicode_age("\u{11F3E}".."\u{11F59}",
                       "KAWI VOWEL SIGN E..KAWI DIGIT NINE")
    assert_unicode_age("\u{1342F}",
                       "EGYPTIAN HIEROGLYPH V011D")
    assert_unicode_age("\u{13439}".."\u{1343F}",
                       "EGYPTIAN HIEROGLYPH INSERT AT MIDDLE..EGYPTIAN HIEROGLYPH END WALLED ENCLOSURE")
    assert_unicode_age("\u{13440}".."\u{13455}",
                       "EGYPTIAN HIEROGLYPH MIRROR HORIZONTALLY..EGYPTIAN HIEROGLYPH MODIFIER DAMAGED")
    assert_unicode_age("\u{1B132}", "HIRAGANA LETTER SMALL KO")
    assert_unicode_age("\u{1B155}", "KATAKANA LETTER SMALL KO")
    assert_unicode_age("\u{1D2C0}".."\u{1D2D3}",
                       "KAKTOVIK NUMERAL ZERO..KAKTOVIK NUMERAL NINETEEN")
    assert_unicode_age("\u{1DF25}".."\u{1DF2A}",
                       "LATIN SMALL LETTER D WITH MID-HEIGHT LEFT HOOK..LATIN SMALL LETTER T WITH MID-HEIGHT LEFT HOOK")
    assert_unicode_age("\u{1E030}".."\u{1E06D}",
                       "MODIFIER LETTER CYRILLIC SMALL A..MODIFIER LETTER CYRILLIC SMALL STRAIGHT U WITH STROKE")
    assert_unicode_age("\u{1E08F}",
                       "COMBINING CYRILLIC SMALL LETTER BYELORUSSIAN-UKRAINIAN I")
    assert_unicode_age("\u{1E4D0}".."\u{1E4F9}",
                       "NAG MUNDARI LETTER O..NAG MUNDARI DIGIT NINE")
    assert_unicode_age("\u{1F6DC}", "WIRELESS")
    assert_unicode_age("\u{1F774}".."\u{1F776}",
                       "LOT OF FORTUNE..LUNAR ECLIPSE")
    assert_unicode_age("\u{1F77B}".."\u{1F77F}",
                       "HAUMEA..ORCUS")
    assert_unicode_age("\u{1F7D9}", "NINE POINTED WHITE STAR")
    assert_unicode_age("\u{1FA75}".."\u{1FA77}",
                       "LIGHT BLUE HEART..PINK HEART")
    assert_unicode_age("\u{1FA87}".."\u{1FA88}",
                       "MARACAS..FLUTE")
    assert_unicode_age("\u{1FAAD}".."\u{1FAAF}",
                       "FOLDING HAND FAN..KHANDA")
    assert_unicode_age("\u{1FABB}".."\u{1FABD}",
                       "HYACINTH..WING")
    assert_unicode_age("\u{1FABF}", "GOOSE")
    assert_unicode_age("\u{1FACE}".."\u{1FACF}",
                       "MOOSE..DONKEY")
    assert_unicode_age("\u{1FADA}".."\u{1FADB}",
                       "GINGER ROOT..PEA POD")
    assert_unicode_age("\u{1FAE8}", "SHAKING FACE")
    assert_unicode_age("\u{1FAF7}".."\u{1FAF8}",
                       "LEFTWARDS PUSHING HAND..RIGHTWARDS PUSHING HAND")
    assert_unicode_age("\u{2B739}",
                       "CJK UNIFIED IDEOGRAPH-2B739")
    assert_unicode_age("\u{31350}".."\u{323AF}",
                       "CJK UNIFIED IDEOGRAPH-31350..CJK UNIFIED IDEOGRAPH-323AF")
  end

  UnicodeAgeRegexps = Hash.new do |h, age|
    h[age] = [/\A\p{age=#{age}}+\z/u, /\A\P{age=#{age}}+\z/u].freeze
  end

  def assert_unicode_age(char, mesg = nil, matches: @matches, unmatches: @unmatches)
    if Range === char
      char = char.to_a.join("")
    end

    matches.each do |age|
      pos, neg = UnicodeAgeRegexps[age]
      assert_match(pos, char, mesg)
      assert_not_match(neg, char, mesg)
    end

    unmatches.each do |age|
      pos, neg = UnicodeAgeRegexps[age]
      assert_not_match(pos, char, mesg)
      assert_match(neg, char, mesg)
    end
  end

  MatchData_A = eval("class MatchData_\u{3042} < MatchData; self; end")

  def test_matchdata
    a = "haystack".match(/hay/)
    b = "haystack".match(/hay/)
    assert_equal(a, b, '[ruby-core:24748]')
    h = {a => 42}
    assert_equal(42, h[b], '[ruby-core:24748]')
=begin
    assert_match(/#<TestRegexp::MatchData_\u{3042}:/, MatchData_A.allocate.inspect)
=end

    h = /^(?<@time>\d+): (?<body>.*)/.match("123456: hoge fuga")
    assert_equal("123456", h["@time"])
    assert_equal("hoge fuga", h["body"])
  end

  def test_regexp_popped
    EnvUtil.suppress_warning do
      assert_nothing_raised { eval("a = 1; /\#{ a }/; a") }
      assert_nothing_raised { eval("a = 1; /\#{ a }/o; a") }
    end
  end

  def test_invalid_fragment
    bug2547 = '[ruby-core:27374]'
    assert_raise(SyntaxError, bug2547) do
      assert_warning(/ignored/) {eval('/#{"\\\\"}y/')}
    end
  end

  def test_dup_warn
    assert_warning(/duplicated/) { Regexp.new('[\u3042\u3043\u3042]') }
    assert_warning(/duplicated/) { Regexp.new('[\u3042\u3043\u3043]') }
    assert_warning(/\A\z/) { Regexp.new('[\u3042\u3044\u3043]') }
    assert_warning(/\A\z/) { Regexp.new('[\u3042\u3045\u3043]') }
    assert_warning(/\A\z/) { Regexp.new('[\u3042\u3045\u3044]') }
    assert_warning(/\A\z/) { Regexp.new('[\u3042\u3045\u3043-\u3044]') }
    assert_warning(/duplicated/) { Regexp.new('[\u3042\u3045\u3042-\u3043]') }
    assert_warning(/duplicated/) { Regexp.new('[\u3042\u3045\u3044-\u3045]') }
    assert_warning(/\A\z/) { Regexp.new('[\u3042\u3046\u3044]') }
    assert_warning(/duplicated/) { Regexp.new('[\u1000-\u2000\u3042-\u3046\u3044]') }
    assert_warning(/duplicated/) { Regexp.new('[\u3044\u3041-\u3047]') }
    assert_warning(/duplicated/) { Regexp.new('[\u3042\u3044\u3046\u3041-\u3047]') }

    bug7471 = '[ruby-core:50344]'
    assert_warning('', bug7471) { Regexp.new('[\D]') =~ "\u3042" }

    bug8151 = '[ruby-core:53649]'
    assert_warning(/\A\z/, bug8151) { Regexp.new('(?:[\u{33}])').to_s }

    assert_warning(%r[/.*/\Z]) { Regexp.new("[\n\n]") }
  end

  def test_property_warn
    assert_in_out_err('-w', 'x=/\p%s/', [], %r"warning: invalid Unicode Property \\p: /\\p%s/")
  end

  def test_invalid_escape_error
    bug3539 = '[ruby-core:31048]'
    error = assert_raise(SyntaxError) {eval('/\x/', nil, bug3539)}
    assert_match(/invalid hex escape/, error.message)
    assert_equal(1, error.message.scan(/.*invalid .*escape.*/i).size, bug3539)
  end

  def test_raw_hyphen_and_tk_char_type_after_range
    bug6853 = '[ruby-core:47115]'
    # use Regexp.new instead of literal to ignore a parser warning.
    re = assert_warning(/without escape/) {Regexp.new('[0-1-\\s]')}
    check(re, [' ', '-'], ['2', 'a'], bug6853)
  end

  def test_error_message_on_failed_conversion
    bug7539 = '[ruby-core:50733]'
    assert_equal false, /x/=== 42
    assert_raise_with_message(TypeError, 'no implicit conversion of Integer into String', bug7539) {
      Regexp.quote(42)
    }
  end

  def test_conditional_expression
    bug8583 = '[ruby-dev:47480] [Bug #8583]'

    conds = {"xy"=>true, "yx"=>true, "xx"=>false, "yy"=>false}
    assert_match_each(/\A((x)|(y))(?(2)y|x)\z/, conds, bug8583)
    assert_match_each(/\A((?<x>x)|(?<y>y))(?(<x>)y|x)\z/, conds, bug8583)

    bug12418 = '[ruby-core:75694] [Bug #12418]'
    assert_raise(RegexpError, bug12418){ Regexp.new('(0?0|(?(5)||)|(?(5)||))?') }
  end

  def test_options_in_look_behind
    assert_nothing_raised {
      assert_match_at("(?<=(?i)ab)cd", "ABcd", [[2,4]])
      assert_match_at("(?<=(?i:ab))cd", "ABcd", [[2,4]])
      assert_match_at("(?<!(?i)ab)cd", "aacd", [[2,4]])
      assert_match_at("(?<!(?i:ab))cd", "aacd", [[2,4]])

      assert_not_match("(?<=(?i)ab)cd", "ABCD")
      assert_not_match("(?<=(?i:ab))cd", "ABCD")
      assert_not_match("(?<!(?i)ab)cd", "ABcd")
      assert_not_match("(?<!(?i:ab))cd", "ABcd")
    }
  end

  def test_quantifier_reduction
    assert_equal('aa', eval('/(a+?)*/').match('aa')[0])
    assert_equal('aa', eval('/(?:a+?)*/').match('aa')[0])

    quantifiers = %w'? * + ?? *? +?'
    quantifiers.product(quantifiers) do |q1, q2|
      EnvUtil.suppress_warning do
        r1 = eval("/(a#{q1})#{q2}/").match('aa')[0]
        r2 = eval("/(?:a#{q1})#{q2}/").match('aa')[0]
        assert_equal(r1, r2)
      end
    end
  end

  def test_once
    pr1 = proc{|i| /#{i}/o}
    assert_equal(/0/, pr1.call(0))
    assert_equal(/0/, pr1.call(1))
    assert_equal(/0/, pr1.call(2))
  end

  def test_once_recursive
    pr2 = proc{|i|
      if i > 0
        /#{pr2.call(i-1).to_s}#{i}/
      else
        //
      end
    }
    assert_equal(/(?-mix:(?-mix:(?-mix:)1)2)3/, pr2.call(3))
  end

  def test_once_multithread
    m = Thread::Mutex.new
    pr3 = proc{|i|
      /#{m.unlock; sleep 0.5; i}/o
    }
    ary = []
    n = 0
    th1 = Thread.new{m.lock; ary << pr3.call(n+=1)}
    th2 = Thread.new{m.lock; ary << pr3.call(n+=1)}
    th1.join; th2.join
    assert_equal([/1/, /1/], ary)
  end

  def test_once_escape
    pr4 = proc{|i|
      catch(:xyzzy){
        /#{throw :xyzzy, i}/o =~ ""
        :ng
      }
    }
    assert_equal(0, pr4.call(0))
    assert_equal(1, pr4.call(1))
  end

  def test_eq_tilde_can_be_overridden
    assert_separately([], <<-RUBY)
      class Regexp
        undef =~
        def =~(str)
          "foo"
        end
      end

      assert_equal("foo", // =~ "")
    RUBY
  end

  def test_invalid_free_at_parse_depth_limit_over
    assert_separately([], "#{<<-"begin;"}\n#{<<-"end;"}")
    begin;
      begin
        require '-test-/regexp'
      rescue LoadError
      else
        bug = '[ruby-core:79624] [Bug #13234]'
        Bug::Regexp.parse_depth_limit = 10
        src = "[" * 100
        3.times do
          assert_raise_with_message(RegexpError, /parse depth limit over/, bug) do
            Regexp.new(src)
          end
        end
      end
    end;
  end

  def test_absent
    assert_equal(0, /(?~(a|c)c)/ =~ "abb")
    assert_equal("abb", $&)

    assert_equal(0, /\/\*((?~\*\/))\*\// =~ "/*abc*def/xyz*/ /* */")
    assert_equal("abc*def/xyz", $1)

    assert_equal(0, /(?~(a)c)/ =~ "abb")
    assert_nil($1)

    assert_equal(0, /(?~(a))/ =~ "")
    assert_nil($1)
  end

  def test_backref_overrun
    assert_raise_with_message(SyntaxError, /invalid backref number/) do
      eval(%["".match(/(())(?<X>)((?(90000)))/)])
    end
  end

  def test_bug18631
    assert_kind_of MatchData, /(?<x>a)(?<x>aa)\k<x>/.match("aaaaa")
    assert_kind_of MatchData, /(?<x>a)(?<x>aa)\k<x>/.match("aaaa")
    assert_kind_of MatchData, /(?<x>a)(?<x>aa)\k<x>/.match("aaaab")
  end

  def test_invalid_group
    assert_separately([], "#{<<-"begin;"}\n#{<<-'end;'}")
    begin;
      assert_raise_with_message(RegexpError, /invalid conditional pattern/) do
        Regexp.new("((?(1)x|x|)x)+")
      end
    end;
  end

  # This assertion is for porting x2() tests in testpy.py of Onigmo.
  def assert_match_at(re, str, positions, msg = nil)
    re = Regexp.new(re) unless re.is_a?(Regexp)

    match = re.match(str)

    assert_not_nil match, message(msg) {
      "Expected #{re.inspect} to match #{str.inspect}"
    }

    if match
      actual_positions = (0...match.size).map { |i|
        [match.begin(i), match.end(i)]
      }

      assert_equal positions, actual_positions, message(msg) {
        "Expected #{re.inspect} to match #{str.inspect} at: #{positions.inspect}"
      }
    end
  end

  def assert_match_each(re, conds, msg = nil)
    errs = conds.select {|str, match| match ^ (re =~ str)}
    msg = message(msg) {
      "Expected #{re.inspect} to\n" +
      errs.map {|str, match| "\t#{'not ' unless match}match #{str.inspect}"}.join(",\n")
    }
    assert_empty(errs, msg)
  end

  def test_s_timeout
    assert_separately([], "#{<<-"begin;"}\n#{<<-'end;'}")
      timeout = #{ EnvUtil.apply_timeout_scale(0.2).inspect }
    begin;
      Regexp.timeout = timeout
      assert_in_delta(timeout, Regexp.timeout, timeout * 2 * Float::EPSILON)

      t = Time.now
      assert_raise_with_message(Regexp::TimeoutError, "regexp match timeout") do
        # A typical ReDoS case
        /^(a*)*\1$/ =~ "a" * 1000000 + "x"
      end
      t = Time.now - t

      assert_in_delta(timeout, t, timeout / 2)
    end;
  end

  def test_s_timeout_corner_cases
    assert_separately([], "#{<<-"begin;"}\n#{<<-'end;'}")
    begin;
      assert_nil(Regexp.timeout)

      # This is just an implementation detail that users should not depend on:
      # If Regexp.timeout is set to a value greater than the value that can be
      # represented in the internal representation of timeout, it uses the
      # maximum value that can be represented.
      Regexp.timeout = Float::INFINITY
      assert_equal(((1<<64)-1) / 1000000000.0, Regexp.timeout)

      Regexp.timeout = 1e300
      assert_equal(((1<<64)-1) / 1000000000.0, Regexp.timeout)

      assert_raise(ArgumentError) { Regexp.timeout = 0 }
      assert_raise(ArgumentError) { Regexp.timeout = -1 }

      Regexp.timeout = nil
      assert_nil(Regexp.timeout)
    end;
  end

  def per_instance_redos_test(global_timeout, per_instance_timeout, expected_timeout)
    assert_separately([], "#{<<-"begin;"}\n#{<<-'end;'}")
      global_timeout = #{ EnvUtil.apply_timeout_scale(global_timeout).inspect }
      per_instance_timeout = #{ (per_instance_timeout ? EnvUtil.apply_timeout_scale(per_instance_timeout) : nil).inspect }
      expected_timeout = #{ EnvUtil.apply_timeout_scale(expected_timeout).inspect }
    begin;
      Regexp.timeout = global_timeout

      re = Regexp.new("^(a*)\\1b?a*$", timeout: per_instance_timeout)
      if per_instance_timeout
        assert_in_delta(per_instance_timeout, re.timeout, per_instance_timeout * 2 * Float::EPSILON)
      else
        assert_nil(re.timeout)
      end

      t = Time.now
      assert_raise_with_message(Regexp::TimeoutError, "regexp match timeout") do
        re =~ "a" * 1000000 + "x"
      end
      t = Time.now - t

      assert_in_delta(expected_timeout, t, expected_timeout * 3 / 4)
    end;
  end

  def test_timeout_shorter_than_global
    omit "timeout test is too unstable on s390x" if RUBY_PLATFORM =~ /s390x/
    per_instance_redos_test(10, 0.2, 0.2)
  end

  def test_timeout_longer_than_global
    omit "timeout test is too unstable on s390x" if RUBY_PLATFORM =~ /s390x/
    per_instance_redos_test(0.01, 0.5, 0.5)
  end

  def test_timeout_nil
    per_instance_redos_test(0.5, nil, 0.5)
  end

  def test_timeout_corner_cases
    assert_separately([], "#{<<-"begin;"}\n#{<<-'end;'}")
    begin;
      assert_nil(//.timeout)

      # This is just an implementation detail that users should not depend on:
      # If Regexp.timeout is set to a value greater than the value that can be
      # represented in the internal representation of timeout, it uses the
      # maximum value that can be represented.
      assert_equal(((1<<64)-1) / 1000000000.0, Regexp.new("foo", timeout: Float::INFINITY).timeout)

      assert_equal(((1<<64)-1) / 1000000000.0, Regexp.new("foo", timeout: 1e300).timeout)

      assert_raise(ArgumentError) { Regexp.new("foo", timeout: 0) }
      assert_raise(ArgumentError) { Regexp.new("foo", timeout: -1) }
    end;
  end

  def test_match_cache_exponential
    assert_separately([], "#{<<-"begin;"}\n#{<<-'end;'}")
      timeout = #{ EnvUtil.apply_timeout_scale(10).inspect }
    begin;
      Regexp.timeout = timeout
      assert_nil(/^(a*)*$/ =~ "a" * 1000000 + "x")
    end;
  end

  def test_match_cache_square
    assert_separately([], "#{<<-"begin;"}\n#{<<-'end;'}")
      timeout = #{ EnvUtil.apply_timeout_scale(10).inspect }
    begin;
      Regexp.timeout = timeout
      assert_nil(/^a*b?a*$/ =~ "a" * 1000000 + "x")
    end;
  end

  def test_match_cache_atomic
    assert_separately([], "#{<<-"begin;"}\n#{<<-'end;'}")
      timeout = #{ EnvUtil.apply_timeout_scale(10).inspect }
    begin;
      Regexp.timeout = timeout
      assert_nil(/^a*?(?>a*a*)$/ =~ "a" * 1000000 + "x")
    end;
  end

  def test_match_cache_atomic_complex
    assert_separately([], "#{<<-"begin;"}\n#{<<-'end;'}")
      timeout = #{ EnvUtil.apply_timeout_scale(10).inspect }
    begin;
      Regexp.timeout = timeout
      assert_nil(/a*(?>a*)ab/ =~ "a" * 1000000 + "b")
    end;
  end

  def test_match_cache_positive_look_ahead
    assert_separately([], "#{<<-"begin;"}\n#{<<-'end;'}")
      timeout = #{ EnvUtil.apply_timeout_scale(10).inspect }
    begin;
       Regexp.timeout = timeout
       assert_nil(/^a*?(?=a*a*)$/ =~ "a" * 1000000 + "x")
    end;
  end

  def test_match_cache_positive_look_ahead_complex
    assert_separately([], "#{<<-"begin;"}\n#{<<-'end;'}")
      timeout = #{ EnvUtil.apply_timeout_scale(10).inspect }
    begin;
       Regexp.timeout = timeout
       assert_equal(/(?:(?=a*)a)*/ =~ "a" * 1000000, 0)
    end;
  end

  def test_match_cache_negative_look_ahead
    assert_separately([], "#{<<-"begin;"}\n#{<<-'end;'}")
      timeout = #{ EnvUtil.apply_timeout_scale(10).inspect }
    begin;
      Regexp.timeout = timeout
      assert_nil(/^a*?(?!a*a*)$/ =~ "a" * 1000000 + "x")
    end;
  end

  def test_match_cache_positive_look_behind
    assert_separately([], "#{<<-"begin;"}\n#{<<-'end;'}")
      timeout = #{ EnvUtil.apply_timeout_scale(10).inspect }
    begin;
      Regexp.timeout = timeout
      assert_nil(/(?<=abc|def)(a|a)*$/ =~ "abc" + "a" * 1000000 + "x")
    end;
  end

  def test_match_cache_negative_look_behind
    assert_separately([], "#{<<-"begin;"}\n#{<<-'end;'}")
    timeout = #{ EnvUtil.apply_timeout_scale(10).inspect }
    begin;
      Regexp.timeout = timeout
      assert_nil(/(?<!x)(a|a)*$/ =~ "a" * 1000000 + "x")
    end;
  end

  def test_cache_opcodes_initialize
    str = 'test1-test2-test3-test4-test_5'
    re = '^([0-9a-zA-Z\-/]*){1,256}$'
    100.times do
      assert !Regexp.new(re).match?(str)
    end
  end

  def test_bug_19273 # [Bug #19273]
    pattern = /(?:(?:-?b)|(?:-?(?:1_?(?:0_?)*)?0))(?::(?:(?:-?b)|(?:-?(?:1_?(?:0_?)*)?0))){0,3}/
    assert_equal("10:0:0".match(pattern)[0], "10:0:0")
  end

  def test_bug_19467 # [Bug #19467]
    assert_separately([], "#{<<-"begin;"}\n#{<<-'end;'}")
      timeout = #{ EnvUtil.apply_timeout_scale(10).inspect }
    begin;
      Regexp.timeout = timeout

      assert_nil(/\A.*a.*z\z/ =~ "a" * 1000000 + "y")
    end;
  end

  def test_bug_19476 # [Bug #19476]
    assert_equal("123456789".match(/(?:x?\dx?){2,10}/)[0], "123456789")
    assert_equal("123456789".match(/(?:x?\dx?){2,}/)[0], "123456789")
  end

  def test_encoding_flags_are_preserved_when_initialized_with_another_regexp
    re = Regexp.new("\u2018hello\u2019".encode("UTF-8"))
    str = "".encode("US-ASCII")

    assert_nothing_raised do
      str.match?(re)
      str.match?(Regexp.new(re))
    end
  end

  def test_bug_19537 # [Bug #19537]
    str = 'aac'
    re = '^([ab]{1,3})(a?)*$'
    100.times do
      assert !Regexp.new(re).match?(str)
    end
  end

  def test_linear_time_p
    assert_send [Regexp, :linear_time?, /a/]
    assert_send [Regexp, :linear_time?, 'a']
    assert_send [Regexp, :linear_time?, 'a', Regexp::IGNORECASE]
    assert_not_send [Regexp, :linear_time?, /(a)\1/]
    assert_not_send [Regexp, :linear_time?, "(a)\\1"]

    assert_not_send [Regexp, :linear_time?, /(?=(a))/]
    assert_not_send [Regexp, :linear_time?, /(?!(a))/]

    assert_raise(TypeError) {Regexp.linear_time?(nil)}
    assert_raise(TypeError) {Regexp.linear_time?(Regexp.allocate)}
  end

  def test_linear_performance
    pre = ->(n) {[Regexp.new("a?" * n + "a" * n), "a" * n]}
    assert_linear_performance([10, 29], pre: pre) do |re, s|
      re =~ s
    end
  end
end