From: "Aleksi Niemelä" <aleksi.niemela@...>
Date: 2000-07-15T14:50:59+02:00
Subject: [ruby-talk:04036] Re: RegExps and latin_1

> Ruby uses isalnum() to determin what is word character, so making the
> extension that calls setlocale(), and calling it before any regexp
> is compiled, may work for you.

I can wait for real i18n, and in the meanwhile use '[\w&auml;]' etc (&auml
is 'a' with umlauts, not five characters :) when needed. One remark: Perl is
not handling this either.

> The REAL internationalization is the theme for 1.7.  Stay tuned.

I can hardly wait :).

One more question: how about allowing redefinitions of rexexp character
classes? I'm not sure how big change this would be in the regexp engine, but
it would surely be neat. Then my (mental) semantic distance would be really
short if I'm coding some text processing utilities for artificial language
'Foo 9 from Outer Space' which has only two characters in whole language:
'f' and '9'. Redefining one word character '\w' to match [f9] would make
things nice.

	- Aleksi