Re: new Apple symbols map (exposes kernel buglets?)
On Saturday 23 August 2003 1:34, Michel Dänzer wrote:
> On Fri, 2003-08-22 at 09:55, Frank Murphy wrote:
> > This should be interesting to anyone with an Apple laptop with the
> > mimized keyboard <apple(laptop)>. It's also mildly interesting to anyone
> > with an Apple desktop keyboard (though it only enables the Help key at
> > the moment) <apple(extended)>.
>
> Can't the Help key simply be fixed in the macintosh symbols? Are there
> Mac keyboards where this is actually an Insert key?
Since Debian 3.0, ppc kernels should generate the same scancodes as i386
kernels (as you know). In X they both use the xfree86 symbols, and I'd like
to increse that convergence. The keycode/key symbol mapping doesn't look
totally consistient (<INS> = Help), but I blame that on the PC-keyboard bias.
:)
As far as an Apple keyboard that has an Insert key beneath the Delete key, I
don't know. Who on this list has an extended keyboard and can share?
> > Unfortunately, I can't seem to configure XFree86 to take this (or any)
> > XkbSymbols option. If someone figures it out, please let me know.
>
> Tell me about figuring out why something doesn't work with Xkb... I had
> to manually strace xkbcomp to find out what was wrong while working on
> the attached patch. Tell me what you think about it, it works with
I haven't tried it yet, but just looking at it, I have a couple of comments.
First, I really like the idea of moving the macos_alt to the altwin map:
much better place. However, I don't want to move the 'extended' or 'laptop'
into a specific macintosh keymap, like you've done here in us. Long-term, I'd
actually like to increase the similarities between PC and Apple keyboards, so
get rid of the macintosh directory and be able to use pc/us(105)+apple or
pc/de+apple to specifiy an Apple keyboard (of course, using the rules so that
configuration is similar to what you have here).
> Option "XkbModel" "powerbook" (or "ibook")
>
> and
>
> Option "XkbOptions" "altwin:macos_alt"
>
> BTW, where do you get the Multi_key definition from? Can't seem to find
> it here.
What do you mean? Where do I define it (in the apple file)? Or where do I know
what the Multi_key does?
> > Also, while trying to create this map, I've run across some buglets
> > (perhaps in the kernel keyboard code) that prevent this map from being
> > complete. I've noted the problems in the file, but basically, the laptop
> > 'Menu' key, and the externded keyboards F13, F14, & F15 keys don't give
> > the expected keycodes. The keycode for the right logo key was broken in
> > kernel 2.4.19, but is fixed in 2.4.21, but I can't find the location of
> > the fix.
>
> You may want to pass this info on to linuxppc-dev and/or an input layer
> list.
OK. I'll post something to the linux-ppc list.
> Speaking of bugs, fn+command generates the same keycode as the eject key
> in X here, but not in console (showkey shows several keycodes for the
> eject key) - maybe related to the keypad equal bug you fixed recently?
I don't have an eject key here, but I imagine it's related to X's KP_Equal
bug. XFree's keyboard handling has a lot of these kinds of problems, but in
95% of the cases it just works, so noone wants to touch it. David Dawes
reacted really quickly, as did Ivan Pascal. David actually made the fix.
Frank
Reply to: