Concerning keyboard mapping,
The PC keyboard sends a scancode when a key is pressed and the
same code+128 when the key is released. This is mapped to the ascii
characters we
expect.
Some terminals, notably the wyse60 in PC-TERM mode can be configured to do
this also.
Obviously this is ideal for DOSEMU.
The catch is that when you first connect to a system the user must present a
username and password,
in ascii, and then enter a recognisable command at the shell prompt.
Scancodes wont be acceptable there.
Or will they?
All the magic of terminal conversion is done in two places.
termcap, which we all know and love,
but also stty, which supports 'raw' and 'cooked' io modes, flow control etc.
In SCO, the stty system can support scancode mode!
>From an admittedly cursory glance, it is possible to use a PC-TERM
terminal to login to SCO as you would use any other terminal.
This then provides an ideal interface to DOSEMU.
AFAIK, the problem then comes back to the stty subsystem and the keymapping.
It should be possible to assign a separate keymap to every tty. (At present
in linux only one 'console' keymap
can be used. It is impossible to assign different keymaps to different
virtual consoles, let alone all the different ttys)
Another possibility is to use wyse60 emulation and throw it in and out of
PC-TERM mode using
control codes. The old ms Word for SCO does this. The catch with this is
that sooner or later
it will get out of sync and also how would you change VC's in that mode?
So really, I think this discussion is in the wrong place. We've got to find
out who maintains
stty and be prepared to do some coding. That lets me out since I'm inept.
Regards
John
