[arch-general] Microsoft keyboard and unworking keys

[André Ramaciotti]

On Wed, Feb 24, 2010 at 9:34 PM, Celti <celticmadman at gmail.com> wrote:

> On Wed, Feb 24, 2010 at 17:32, André Ramaciotti
> <andre.ramaciotti at gmail.com> wrote:
> > On Wed, Feb 24, 2010 at 7:45 PM, Andre Ramaciotti <
> > andre.ramaciotti at gmail.com> wrote:
> >
> >> Celti <celticmadman at gmail.com> writes:
> >>
> >> > On Wed, Feb 24, 2010 at 15:38, Andre Ramaciotti
> >> > <andre.ramaciotti at gmail.com> wrote:
> >> >> Celti <celticmadman at gmail.com> writes:
> >> >>
> >> >>> On Wed, Feb 24, 2010 at 15:23, Celti <celticmadman at gmail.com>
> wrote:
> >> >>>> On Wed, Feb 24, 2010 at 15:19, Andre Ramaciotti
> >> >>>> <andre.ramaciotti at gmail.com> wrote:
> >> >>>>> Daenyth Blank <daenyth+arch at gmail.com <daenyth%2Barch at gmail.com><
> daenyth%2Barch at gmail.com <daenyth%252Barch at gmail.com>>>
> >> writes:
> >> >>>>>> On Wed, Feb 24, 2010 at 16:34, Andre Ramaciotti
> >> >>>>>> <andre.ramaciotti at gmail.com> wrote:
> >> >>>>>>> as in "they don't show any keycode
> >> >>>>>>> on 'xev' or 'showkey'".
> >> >>>>>>
> >> >>>>>> It sounds like those keys are broken. It should be sending
> >> something. Is it old?
> >> >>>>>
> >> >>>>> It's barely used.  I don't think the keys are broken, most
> probably,
> >> >>>>> these keys send the "I've been pressed" signal in a non-standard
> way
> >> >>>>> (you know, it's Microsoft, it wouldn't surprise me at all).  I
> >> suspect
> >> >>>>> this because this keyboard comes with an installation CD, which
> >> probably
> >> >>>>> contains some kind of special driver (for Windows and Mac OS X
> only).
> >> >>>>>
> >> >>>>
> >> >>>> Likely they don't have a kernel mapping, so X doesn't even see
> them.
> >> >>>> You'll need to get their scancodes with `showkeys`, and map them to
> >> >>>> keycodes with `setkeycodes`, while out of X.
> >> >>>>
> >> >>>> ~celti
> >> >>>
> >> >>> Er, sorry, I'm blind. You said you used showkey. Did you try it with
> >> '-s'?
> >> >>
> >> >> Yay! It did return some key codes, though they were kind of strange,
> >> >> like a single key printing '0xe0 0x5d' (instead of a single byte).
> Will
> >> >> I have any problems because of this?
> >> >
> >> > Nope. Those are scancodes, not keycodes. There are plenty of howtos
> >> > out there on mapping scancodes to keycodes; it's been long enough
> >> > since I've needed to that I've forgotten the exact syntax.
> >> >
> >> > Glad I could help.
> >> >
> >> > ~celti
> >>
> >> I'm glad, too. Thank you! :)
> >>
> >
> > Actually... I should pay more atention to what I'm doing.  I've just
> > realized that I was pressing the wrong key; I was pressing one by the
> side
> > of the key that doesn't produces any code.  They key I should be pressing
> > still doesn't show anycode when running showkey as `showkey -s`.
>
> Oh. In that case, it probably does use some silly proprietary
> Microsoft thing. No idea how to help there, sorry.
>
> ~celti
>

I've found something interestingly weird. The output of ls -lh
/dev/input/by-id shows that this keyboard has three /dev/input entries:
/dev/input/event14, /dev/input/event15 and /dev/input/js0 (?) . Event14 is
the "main keyboard", where most of the keypresses result in an event. Both
event15 and js0 react to only one key (one of the non-working ones). So
mistery somewhat solved: it is indeed a Microsoft thing, and I'm
semi-officially giving up making this keyboard work 100%. I has more
multimedia keys that I'll ever need, anyway (most of them working).