Video and Keymap Quirks
Richard Hughes
hughsient at gmail.com
Tue Jul 10 12:41:14 UTC 2007
On Tue, 2007-07-10 at 14:35 +0200, Hans de Goede wrote:
> Richard Hughes wrote:
> > On Tue, 2007-07-10 at 13:31 +0200, Hans de Goede wrote:
>
> >> I've been working on this, but currently I'm stuck on getting my intermediate
> >> solution for the generation of XKeysyms for these keys integrated into the
> >> xkeyboard-config package. I say intermediate, because in the long run X should
> >> switch from the kbd driver to the evdev driver, changing the way keyysyms get
> >> generated significantly.
> >
> > Sure, that will be done before F8.
> >
>
> Will it, is evdev scheduled for F-8, this the first time I've heard this. Also
> evdev upstream has told me that evdev currently cannot handle chinese/japanese
> input, so I think a switch atm is very unlikely.
Ahh, I didn't know this. Is this a long term problem or just a bug?
> I'm mainly focusing on normal (ps2/usb keyboards) with extra keys for
> multimedia and internet stuff. My solution currently assumes that:
> 1) The kernel generates the correct keycodes for these keys (which is
> true for all usb keyboards and most ps/2 keyboards)
> For the few exceptional ps/2 non laptop keyboards manual setkeycodes calls
> will be necessary, but that will be the case no matter what so for now I'm
> ignoring this.
Cool.
> 2) That F-8 will keep using the kbd driver for Xorg, if F-8 will change to
> evdev, then my plan needs to be reworked.
Right. Which may or may not be true.
> With this assumptions and the focus of internet / multimedia keys, my simple
> and short plan is to:
> 1) Add Linux virtual ps/2 keyboard fake scancode -> keysym translation to the
> (default) "Generic 104 keys" and "Generic 105 keys (international)" keyboard
> models. This is what the patch in bug 245576 does.
Yes, that seems sane.
> 2) Start patching applications (starting with metacity) to have default
> keybindings to these keysyms for relevant actions. Currently atleast for
> metacity the default keybindings for multia media actions are empty.
Sure, also sane.
> Notice that 2) is something which will also be necessary once switched to
> evdev, so the only work which will be thrown away because of me choosing in
> intermediate solution (in order to have something working for F-8) is the
> (already written) xkb patch
Then why don't we just merge this now? Could it break anything that
works already?
Richard.
More information about the devel
mailing list