[immodule-qt] Does QT 4 beta 1 support IM modules

[Lars Knoll]

On Saturday 01 January 2005 18:31, Daisuke Kameda wrote:
> Lars Knoll <lars at trolltech.com> wrote:
> > > Thank you for report. But, I was disappointed about its contents.
> >
> > I have for the moment chosen not to add the switcher to the Qt
> > distribution. We have thought about this and would like to integrate the
> > switching framework into Qt.
> >
> > The reason for this is that we need to find a general solution to the
> > problem of switching input languages. This does also include simple X11
> > keyboard layouts and is not only limited to input methods. This is why a
> > switcher plugin won't work.
>
> Since I cannot fully understand the reason which you decided so by this
> explanation, please explain to me a little more concretely.

For the regular user an input method is just another keyboard layout (although 
one that is significantly more complex to use) and he'd like to be able to 
use it as such. I'd like to be able to configure my desktop to e.g. have a 
german, english and japanese "keyboard layouts". As a user I don't care 
technically one of them is an input method and the others are X11 keyboard 
layouts.

In the example above this means that I would like to be able to siwtch between 
these three layouts using the context menu on a lineedit. Your approach with 
the im switcher only allows me to choose japanese in there, but not to switch 
back to german.

This means that I would like to see a unified keyboard switching mechanism 
available on X11 (KDE and Gnome) that we can use. The mechanism should work 
similar to what Windows offers and not make a distinction between regular 
keyboard layouts and input methods.

Once we have a protocol in place, Qt probably needs to do a bit more (or some 
other work as well) than just switching QInputContext's, esp when swithcing 
between regular keyboard layouts and input contexts.

> And, it is possible to make the method of changing input method using D-BUS
> as more general mechanism.

I don't think dbus is the correct solution for this, as it is limited to one 
machine and wouldn't work transparently on the network. In my opinion we will 
need some X11 mechanism to achieve this.

> > > If the reason a part of functions are not adopted, and other part are
> > > remaining exists, please let me know, Lars. Unless the present status
> > > and the future plan in Trolltech are known, we can't decide our future
> > > work.
> >
> > My idea is to have the framework needed to load and work with the IME
> > plugins inside Qt. Most plugins for the different input methods should
> > probably continue to be developed by you (e.g. on freedesktop.org).
>
> I think that QMultiInputContext and QSimpleInputContext as structure which
> a input method can delegate key event to other are absolutely required.
> It will become  mechanism inferior to immodule of gtk+, if it is not
> realized.

I can understand that we should add the simple input method. But the multi 
input context is just a switcher and for the reasons I told you above I would 
like this to go directly into Qt so we can unify this with general keyboard 
layout switching in the long term. In the short term, I would like to add 
some mechanism similar to what QMultiInputContext currently does into Qt (ie. 
switching between input methods) and add configuration options for it to 
qtconfig.

Cheers,
Lars