Re: About the charset and unicode manipulation functions

[Matthias Clasen <mclasen redhat com>]

On Fri, 2005-12-16 at 07:39 +0800, Qiu, Michael wrote:
> I’m working on LSB-desktop working group.
> 
>   I’m responsible for glib test cases development. I encounter some
> problems about the Charset and Unicode manipulation.
> 
>   I develop some test cases for the functions g_locale_from_utf8 and
> g_utf8_collate. For these two functions, I use the zh_CN.GB2312
> encoding.
> 
>   But, when I test these cases on Mandriva and Ubuntu, I find I have
> to install this encoding first, or else the results are failing.
> 
>   Could you please tell me which methods I can use to make these test
> cases independent on the locale?
> 
>   The attachments are the snippet from the test suite.
> 

It is hard to test locale-dependent functions without relying on
installed locales... the GLib unit tests have the same problem. 
You can consult the existing unit tests for g_utf8_collate() and
g_convert() in tests/unicode-collate.c and tests/convert-test.c

Matthias