[SCIM] Re: gtk IM & reset ...

Zhe Su james.su at gmail.com
Mon Dec 13 09:46:10 PST 2004


Hi,
  Sounds reasonable. I have sent a mail to m17n list to suggest them
implement such feature.

Regards
James Su


On Mon, 13 Dec 2004 12:43:28 -0500, Dan Williams <dcbw at redhat.com> wrote:
> On Tue, 2004-12-14 at 01:14 +0800, Zhe Su wrote:
> 
> 
> > Hi,
> >   But the truth is that not all imengines have the abliity to reset
> > themselves. For example m17n lacks such features to reset an input
> > context.
> >   So I think the best way is to do not assume that
> > gtk_im_context_reset () will always success. If the input context can
> > be reset, then it will fire the preedit change signal to OOo to reset
> > the preedit string.
> >   So please do not reset the preedit string in OOo unless it got reset
> > by input context explicitly.
> 
> The problem with this is that there's no other way to reset the preedit
> string.  For example, when you're in OOo Calc, and you wish to use the
> Tab key to move between cells, when the user hits that key, we need to
> tell the imengine/LE to reset itself and clear out the preedit buffer.
> The LE has no way of knowing that we've moved to another cell and that
> the preedit string needs to be reset, so OOo has to do that with
> gtk_im_context_reset().  gtk_im_context_reset() is greatly needed
> functionality from a user's perspective so that things behave correctly.
> 
> If the IM doesn't support it, then the IM needs to implement
> gtk_im_context_reset().
> 
> Email from Owen Taylor:
> ---------------
> > And that no IM I know of actually implements gtk_im_context_reset()...
> > (at least the IIIMF guys say theirs doesn't).
> 
> ???
> 
> GtkIMContextXIM, GtkIMContextSimple certainly do. I can't imagine an
> input method working very well without it.
> 
> Regards,
>                                                 Owen
> ---------------
> 
> Owen certainly believes that input methods need to support
> gtk_im_context_reset(), and Owen is God :).  What keeps m17n from being
> able to do so?
> 
> Dan
> 
>


More information about the scim mailing list