Re: [gtk-list] Re: gdkkeysyms.h question
- From: Andreas Leitner <leitner eiffel com>
- To: gtk-list redhat com
- Subject: Re: [gtk-list] Re: gdkkeysyms.h question
- Date: Fri, 17 Sep 1999 14:08:23 -0700
Havoc Pennington wrote:
> On Fri, 17 Sep 1999, Andreas Leitner wrote:
> > I have no idea about any road map for GTK, but do you consider that this
> > additional abstraction is a valuable goal for one of the next versions?
> It might be reasonable to mash all the Tab variants into GDK_Tab,
> that makes sense to me. I don't make the decisions though. :-)
> We did see a keyboard with a separate tab key on the keypad today at the
> computer store, so go figure. Maybe someone out there would like to have
> that key do something different from the regular tab key. It does make me
> a little nervous to keep them from doing that.
I see. But my suggestion was not only for tab, but for all those other
common keys CAPS-LOCK, SHIFT, ...
>From my point of view, things like GLIB or even XLib should hide details
like that from the customer. The Xlib keyboard routines are there, so I
do not need to programm for every keyboard out there, but only for one
(abstract) keyboard. Of course one has to think hard to find a good
abstraction, that does not prevent using all the features of some
esoteric keyboards. But the current apporach seems to me pretty much
like a hack.
If there is a keyboard that has a 2 different tab keys, we should find a
way to distinguish those keys for those applications that really need to
take care of that fact. But we must also find a clean and easy way for
all the other 99.9% (; of applications to deal with just TAB keys
(mostly they don't care which TAB has been pressed, they are only
interested in the fact that TAB _has_ been pressed).
One way to do that would be:
Declare a certain set of constants from the keysyms file as the basic
set. This set must not have redundant keys. Then provide a function that
takes a keyval and if the given value is not in the basic set, return
the corresponding value from the basic set.
Ie (pseudo code):
int translate_keysm (int keyval)
// translate keyval into one of the keys in basic set
For example: if we declare a certain set of constants from the keysyms
table as standard, and the provide a function that takes a keyval and
converts it for all values that are not in the basic set into a value
from the basic set.
This way applications _may_ choose between all the information or the
] [Thread Prev