Re: Protocol wskbd vs. Protocol standard

[Izumi Tsutsui <tsutsui%ceres.dti.ne.jp@localhost>]

macallan@ wrote:

> on Monday 29 July 2013 13:19:17 Izumi Tsutsui wrote:
> > macallan@ wrote:
> > > For some of them there are translation maps in the xf86-input-kbd driver
> > > so no such workarounds should be necessary in most cases ( mostly ADB and
> > > Sun keyboards that is ).
> >
> > Well, it depends on the definition of "most cases."
> 
> I'm going by mine of course ;)
> ( sorry I got you wrong )

No problem, probably we should split Tier-II ports into Major/Minor...

> > I meant hp300, luna68k, newsmips, ews4800mips, ...
> > (I have not tried /usr/X11R7/lib/X11/xkb/keymap/sony for newsmips though)
> 
> Yeah, I didn't think of those but macppc, sparc, sparc64 etc.

Note sparc64 is still Tier-I :-)

> > mkbd(4) on dreamcast has the similar map to USB but they are
> > not quite identical, so some keys ('\' etc) are not usable.
> 
> That shouldn't be too hard to fix, or am I missing something?

There is no technical difficulties, but it's a bit boring
to implement translation tables and requires a certain motivations..
(an OpenBSD/luna88k guy implemented COMPAT_RAWKBD rawmaps for
 LUNA-88K2 to demonstrate it at Open Source Conference 2013 Nagoya :-)
 http://movapic.com/ebijun/pic/4958994
 http://twitpic.com/cygxab

Anyway we need a consensus of our design goal, and
I also wonder if we can generate translation maps automatically
(or prepare maps dynamically at runtime) by WSKBDIO_GETMAP ioctl.

The old XFree Xhpc server (for hpcmips/hpcarm/hpcsh) seems to
have some piece of code to "generate keymap on the fly"
http://mail-index.netbsd.org/source-changes/2004/07/22/0026.html
but I'm not sure if we can still use this strategy on modern Xorg
servers.

---
Izumi Tsutsui