Bug #7501: \w in a regular expression doesn't match international characters

• Status: Rejected
• Assignee: -
• Target version: -
• ruby -v: ruby 1.9.3p0 (2011-10-30 revision 33570) [i686-linux]

[eltomito (Tomas Partl)]

When using regexp matching, \w doesn't match characters which are not in the English alphabet.
For example, the characters "žščřďťňaáéíóůúý" should all be matched by \w but aren't.

This program demonstrates the bug:

---

encoding: utf-8

match = /\w+/.match( "abcdefghijklmnopqrstuvwxyz" )
puts match.to_s

match = /\w+/.match( "áéíóůúýžščřďťň" ) #some Czech characters
puts match.to_s

match = /\w+/.match( "üäö" ) #some German characters
puts match.to_s

Expected output:

abcdefghijklmnopqrstuvwxyz
áéíóůúýžščřďťň
üäö

Actual output:

abcdefghijklmnopqrstuvwxyz

---

/[[:alpha:]]+/ should behave as you expect

[shyouhei (Shyouhei Urabe)]

• Status changed from Open to Rejected

If I remember correctly this is an intentional design. Because as Unicode version grows, the definition of what is a word character and what is not changes form time to time. It is hard for us to follow that.