From: brabuhr@...
Date: 2010-06-18T01:04:28+09:00
Subject: Re: nothing new in ruby_core for four days ?

>>> But
>>> 1.9.2 implies (or is it 1.9.x ?) - alas poor Yorick! -  changing
>>> Regexp's : \d, \s,\w and the like work no more with utf-8 encodings!
>>
>> Care to share an example, or point to a blog with more details? I have
>> an older 1.9.2 lying around, it doesn't seem to behave how you describe.
>
> http://svn.ruby-lang.org/repos/ruby/tags/v1_9_2_preview3/NEWS
> * \d, \s, and \w are now ASCII only; use POSIX bracket classes and \p{} for
>  Unicode semantics
>
> http://redmine.ruby-lang.org/issues/show/3376
> The rationale is that many people expects \d and \s to match ASCII only.

http://blade.nagaokaut.ac.jp/cgi-bin/scat.rb/ruby/ruby-core/30543
For your convenience, I tell you two well-known incompatibilities:

- \d, \s and \w in Regexp does match *only* ASCII characters.
  Especially, \w does NOT match with accented character, umlaut,
  etc.  In other words, it behaves the same as 1.8.

    # coding: UTF-8
    p(/\w/ =~ "a")  #=> nil

  If you want Unicode sensitive character class, you can use
  \p{Digit} instead of \d, \p{Space} instead of \s, and \p{Word}
  or [\w\P{ASCII}] instead of \w.