Backport #8650
closed
Unexpected result of Regexp#to_s with utf-16 and utf-32 string.
Description
I found the result of Regexp#to_s is incorrect with utf-16 and utf-32 encoded string.
C:\Users\phasis>irb
irb(main):001:0> Regexp.new('abcd'.encode('UTF-16LE'))
=> /a b c d /
irb(main):002:0> Regexp.new('abcd'.encode('UTF-16LE')).to_s
=> "\u3F28\u6D2D\u7869\u613A\u6200\u6300\u6400\u2900"
irb(main):003:0> Regexp.new('abcd'.encode('UTF-16BE'))
=> / a b c d/
irb(main):004:0> Regexp.new('abcd'.encode('UTF-16BE')).to_s
=> "\u283F\u2D6D\u6978\u3A00\u6100\u6200\u6300\u6429"
irb(main):005:0> Regexp.new('abcd'.encode('UTF-32LE'))
=> /a b c d /
irb(main):006:0> Regexp.new('abcd'.encode('UTF-32LE')).to_s
=> "\u{6D2D3F28}\u{613A7869}\u{62000000}\u{63000000}\u{64000000}\u{29000000}"
irb(main):007:0> Regexp.new('abcd'.encode('UTF-32BE'))
=> / a b c d/
irb(main):008:0> Regexp.new('abcd'.encode('UTF-32BE')).to_s
=> "\u{283F2D6D}\u{69783A00}\u6100\u6200\u6300\u6429"
Same result for Ruby 1.9.3 and Ruby 2.0.0
Updated by 
Updated by