From: Tanaka Akira <akr@...>
Date: 2008-01-01T17:07:07+09:00
Subject: Re: [Spam] Re: Wide strings to ruby strings

In article <6.0.0.20.2.20071231185558.0b1886f0@localhost>,
  Martin Duerst <duerst@it.aoyama.ac.jp> writes:

> This may be the best option for cases such as Solaris,
> but it risks a serious hickup or data loss if wcs is UTF-16
> and mbs is not UTF-8.

What I'm not sure is how many application operate Unicode
wchar_t directly.

If an application doesn't operate Unicode wchar_t, wchar_t
has characters only converted from mbs.  So even if wchar_t
is Unicode and mbs is US-ASCII, there is no problem.  All
wchar_t value in the application can be converted to
US-ASCII.

I agree that the problem arise if an application operate
wchar_t directly as Unicode.  Do you know an example of such
application?
-- 
Tanaka Akira