From: Damien Wyart Date: 2002-02-14T06:14:58+09:00 Subject: Re: Regexps and latin-1 * GOTO Kentaro in comp.lang.ruby: > freq.rb uses /\w+/ equivalent to /[0-9A-Za-z_]+/. > Latin-1 words matches to /[\w�-�]+/. OK, this works, but the sort done afterwards is not correct. In French, one should have e < � < f, and freq.rb (with your proposed modification) doesn't lead to this ordering. Any (simple) suggestions ? Sorry, I am not (yet) a Ruby expert... Many thanks, -- Damien Wyart