Keyboard mapping

[Owen Taylor]

On Sun, 2005-05-22 at 17:29 -0400, Jon Smirl wrote:
> On 5/22/05, Owen Taylor <otaylor at redhat.com> wrote:
> > Remember, the end goal for Xgl isn't "X for console nuts" (no offense
> > intended for any such reading this), but rather to replace the XAA-based
> > X server.
> > 
> > As such, XKB support and compatibility with the existing solutions
> > people are using is much more important than matching the kernel's
> > keymap files.
> 
> I would like to make Xgl multiuser from the beginning. To do this
> several problems need to addressed. I'm sure the Ruby people can fill
> in anything I missed. 

While I think putting too much emphasis on this too early could
be a mistake (what percentage of systems are we talking here?),
if it comes naturally, it certainly is nice to have.

[ ... deleting all discussion of non-keyboard stuff ... ]

> 
> That leaves keyboard. The right solution here might be to go with the
> X maps and work towards eliminating the in-kernel one. It is certainly
> possible to implement something that looks just like kernel VCs but
> that runs in user space. Running in user space means it can use
> libraries in common with XGL.

I just don't see any relevance of multiuser to keyboard maps. In the
end, we're running an X server on one of the heads, we present a 
keymap to the user. The keymap is something that the *user* can
configure within X, for themselves. We already have a fancy 
mechanism for doing this within X, XKB. All that is needed to 
implement this is access to the hardware keycodes.

Trying to adapt the current "nested" Xgl keyboard code that is designed
to run on another server with a keymap already in place isn't the right
approach for Xgl owning the keyboard, except possibly as
a temporary stop-gap.

Regards,
						Owen

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://lists.freedesktop.org/archives/dri-egl/attachments/20050522/7a6ccd13/attachment.pgp