[Uim] Re: Suggestion: implement uim_get_language_name_from_locale() by Scheme

[YamaKen]

At Wed, 6 Oct 2004 12:36:30 +0900,
tkng at xem.jp wrote:
> 
> On Mon, 04 Oct 2004 10:37:06 +0900
> YamaKen <yamaken at bp.iij4u.or.jp> wrote:
> 
> > Oops, I had misexpressed my intention. The langcode -> langname
> > translation feature (i.e. lang-code->lang-name procedure) in
> > Scheme world is the flexibility we will get.
> 
>  I see. In your opinion, existence of angcode->langname function in
> scheme world is important, there's no problem even if entity of the
> function is written in C.
> 
> I propose that we should keep uim_get_language_name_from_locale() in C
> and make scheme binding of the function by init_subr, from the point of
> view of memory consumption.

It seems to most benefical for us in this case. Okay, I'll
implement it so.  It reduces not only memory consumption but a
programming cost (i.e. it does not require free()).

Currently I'm going to implement it as follows. Stop me if you
feel something wrong.

- gettext() must be called outside of libuim (as in original
  implementation)

- Reorganize the elements in iso-639-1.def as following if no
  intentional reason exists.

  * current definition

   {N_("Abkhazian"), "ab"},
   {N_("Afar"), "aa"},
   {N_("Afrikaans"), "af"},

  * my preference

   {"ab", N_("Abkhazian")},
   {"aa", N_("Afar")},
   {"af", N_("Afrikaans")},

-------------------------------
YamaKen  yamaken at bp.iij4u.or.jp