> ok. Do you think that this accumulation is something gii should handle ?
> I mean, it would certainly be clean to just expose the unicode once it is
> completed. Then the client would only see one single key press even though
> there were quite a couple.
Though I agree it would be the cleanest solution, the menu stuff would
require heavy interaction with the main application to avoid chaos to break
out on screen.
> since this seems indeed to be very complex, might I suggest that we handle
> this inside berlin ? This way, we could define ourself helper applets which
> present contextual help for composing text. This would be extremely
> difficult on a lower level.
Yes - the main point is the display of completion menus and stuff. This is
very hard to handle on lower levels, especially if you allow multithreading.
We'd basically have to lock areas for the menus, which essentially can't be
done without application help and it would as well couple at least LibGII
to LibGGI, which isn't a good thing either.
As (according to the other post here) there seems to be a Japanese input
mapper available, it would probably be best, if LibGII just reported the
unicode values on the keyboard and left the upper layers to Berlin that
would use said mapper internally.
= Andreas Beck | Email : <[EMAIL PROTECTED]> =