From: Marcus Sundberg <e94_msu@elixir.e.kth.se>
  To  : ggi-develop@eskimo.com
  Date: Thu, 13 Aug 1998 04:04:21 +0200

Re: ggi_key_event confusion

What I mean is that we need an API to recieve keyboard events.

What we currently have is not an API, it's an include file with
#defines and a bunch of targets that translates native keyboard
events into those defines in a mostly random way.

And as I said we need to have exactly one entry describing the
physical key, exactly one entry describing what it means, and
exactly one value corresponding to each key.


> > What we need is one entry describing what physical key was pressed
> > and one entry describing the meaning of that key, with regard taken
> > to modifiers and such. And we need _every_ supported value for both
> > these entries to have a corresponding GGI_* define.
> 
> ???
> 
> That is the purpose of label.  code is platform (i.e. keyboard) dependent.

Well, label is a highly braindead name, because the key to the left of
backspace has a different label on your keyboard than on mine. ;)
In that case I suggest we remove label and make code have that
purpose.

> > I suggest that we have a "code" entry with corresponding GGI_KCODE_*
> > defines and a "sym" entry with corresponding GGI_KSYM_* entries.
> 
> Or you mean we should have separate defines for the key's label and the
> sym?

Yes, that is one thing we definitely should have.
The sym should for ascii keys have the keys asciivalue for example, 
while the code (what you call label) should just be a random number,
like the SCANCODE_* defines in SVGAlib.

//Marcus