[immodule-qt] New unified patch has been released as stable

[Daisuke Kameda]

On Tuesday 07 September 2004 01:04, YamaKen wrote:
> > - It will be better to include QXIMInputContext and QMultiInputContext in
> > libqt.
>
> I don't think so.
>
> From the viewpoint of developer, it will cost us special
> handling in several place. It's a most significant reason for
> me.
>
> In addition, I think that the two implementations should be
> isolated from Qt sourcecode to keep cleaner immodule
> implementation in Qt. If the two are combined into libqt, easy
> layer violation may pollute libqt implementation.
>
> > Now, all sublass of QInputContext corresponding to input method is
> > served as plugin. Therefore, XIM and imsw-multi may be unable
> > to be used when the package of plugin is another and not installed.
>
> I think that it's a responsibility of packager or
> distributor. If a packager want to avoid such happenings, the
> two plugins should be included in libqt package. On the other
> hand, if another distributor want to exclude XIM plugin or
> imsw-multi, it should be able to be excluded.

Your opinion is good in Qt4. But, in Qt3, making QXIMInputContext
into plugin is big change about using XIM. It may be necessary to 
change the policy of packaging.

As other examples, when a user only performs "make sub-src", 
Qt application cannot use XIM with the library. Is this natural for a user?

You may think that there is hardly the users who compile Qt by 
themselves. But, as long as the BBS and ML of a user group to 
which I belong (Japan KDE Users' Group) are seen, the number 
of such users is not few.


-- 
Daisuke Kameda <kaminmat at cc.rim.or.jp>