[Uim] traditional Chinese py input?

[TOKUNAGA Hiroyuki]

On Mon, 19 Jan 2004 13:12:20 +0100
David Oftedal <david at start.no> wrote:

> 
> > I apologize for bringing up so many problems.
> 
> I don't think we're dealing with a lot of problems though.
> 
> The problems with loader.scm could be caused by a lot of things, but 
> remember that you have to add your input method to at least one file
> in addition to that one. One that I can immediately remember is 
> /etc/gtk-2.0/gtk.immodules. There might be others in the
> /usr/share/uim directory as well. Try to grep -Hi "PY.scm" * and see
> for yourself.

Thanks, I forgot to explain them. Currently, input method of uim is
dealed as independent input method from Gtk+.(Although there are a little
defects in this expression, but I can't write better expression.)

So, to add new input method to your system, you must update your
/etc/gtk-2.0/gtk.immodules. I think no file must be updated except
/etc/gtk-2.0/gtk.immodules.


> As for the caching of candidates, you might be right. I've noticed how
> uim-anthy remembers what kanji I select for a particular word for at 
> least the duration of  session, even though there's no backend server 
> there or anything like that. But I guess that's just another reason to
> rename your file instead of calling it PY.scm.

FYI, Anthy learn selected candidate and save them to ~/.anthy/. And if
you want to regiter new word, you can use sumika, a management utility
for dictionaries for input methods.

You can download sumika from http://sourceforge.jp/projects/sumika/, but
if you are using debian or gentoo, you can use apt or emerge to install
sumika.

And, uim-py doesn't learn selected candidates, because uim-py doesn't
have functions to learn selected candadates.


Regards,

-- 
TOKUNAGA Hiroyuki