[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-devel
Subject:    Re: KDE XIM Server
From:       Tim Smith <tim () electronghost ! co ! uk>
Date:       2002-04-19 21:21:34
[Download RAW message or body]

On Thursday 18 Apr 2002 7:51 am, Justin scribed numinously:"
> > > Someone needs to figure out XIM first though..
> >
> > Yes. I'm not sure if the advantages of a new Qt XIM-based system are
> > worth the considerable effort involved in writing one, though.
>
> From what we've seen in the kinput2 code, XIM itself does not look like a
> very complicated protocol.  The trouble is that kinput2 is riddled with
> references to Xt, which makes the code seem 10 times more confusing than
> it should be. If I weren't busy with so many other projects, I'd probably
> invest more time into this and figure it out.
>
> > (Does your input method use cannaserver?)
>
> No.  IMO, the idea of a Cannaserver is yet another unnecessary
> complication. Our Japanese IME just loads the whole dictionary into
> memory (a few MB, which is easy to come by these days), and we use QDict
> to perform fast lookups.

I'm going to have a play this weekend and see if progress outweighs 
frustration. I've been looking over the XIM docs, which seem to define a 
huge amount of stuff nobody in their right mind would ever use. I get the 
same "too heavyweight for my purposes" feeling from the li18nux  stuff. XIM 
seems more widespread, and an XIM server would support stuff like 
OpenOffice too.

Is the code from your IME available anywhere for me to look over? I'm 
thinking more in terms of parsing a flat file into a decision tree on load 
for data structures, but the more XIM stuff I can find the better, 
particularly if you've already got a class for it :-)

-- 
Tim Smith (tim@electronghost.co.uk)
It's the frost. Sometimes it makes the blade stick.

 
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<
[prev in list] [next in list] [prev in thread] [next in thread]