JRuby should support matching "grapheme clusters" (glyphs), which are constructed using mutliple Unicode codepoints.

Expected Behavior (MRI)

glyphs = "\u{61 308 62}".scan(/\X/) # => ["ä", "b"]`
glyphs.map{ |e| e.codepoints.map{ |f| f.to_s(16) } } #=> [["61", "308"], ["62"]]

Actual Behavior (JRuby)

glyphs = "\u{61 308 62}".scan(/\X/) # =>  ["a", "b"]`
glyphs.map{ |e| e.codepoints.map{ |f| f.to_s(16) } } #=> [["61"], ["62"]]

Related Links

• Unicode® Standard Annex Patch for String#encode(opts) [JRUBY-5633] #29: Unicode Text Segmentation: http://unicode.org/reports/tr29/
• Introduction of \X to Onigmo: k-takata/Onigmo@dde0c43
  • Non-Unicode should match: (?>\x0D\x0A|(?m:.))
• Seems to only work properly in MRI 2.4: Remove AS::Multibyte's unicode table rails/rails#26743
• Related Ruby issue: https://bugs.ruby-lang.org/issues/12831
• Example graphemes: https://github.com/janlelis/uniscribe/blob/master/spec/uniscribe_spec.rb