VGA X11.4 server problem

[Erik Fortune]

In article <201 at rebma.rebma.mn.org>, wmo at rebma.rebma.mn.org (Bill
O'Hanlon) writes:
> I'm a little curious if the bit about the RT keyboard is correct...
>	1) the bit about the RT keyboard--the code in that place
>	   mentions the regular PC keyboard as being the other
>	   possibility.  I can't figure out why I'm getting RT there, 
>	   unless I'm s'posed to.

I haven't looked at the code that was recently posted, but I suspect that
much of it was borrowed from the original IBM code.  First some history on
that server, then a long answer (short answer -- the message isn't a problem).

The IBM VGA server ran (originally) on an RT coprocessor card that fit 
into a PS/2; the coprocessor ran BSD un*x.   This combo (PS/2+coprocessor) 
was known as a 6152 and was a *nice* little machine.    The server on 
this machine was *the same binary* as the server on the RT.   It noticed
where it was running and used the correct screens.   The RT was the first
machine to use the new "converged" keyboard layout, which had the caps
lock featured real prominent-like, an the control key strategically hidden
way down in the corner.   We didn't like this, so we swapped (in software)
the Caps Lock and Control keys.   PCs of this era had accesible control
keys, so we called the option to put the control key somewhere useful "-pckeys"
and the option to use the keys as labelled "-rtkeys."

Eventually, PC keyboards converged and more and more of our users wanted
the control key way down there.  We added an informational message to tell
people which keyboard configuration they were getting so there'd be
no surprises.

Your message ("using RT keyboard") really means "not swapping control and
caps lock."   The code to deal with the keyboard was fairly OS-specific,
so I have no idea if the key to actually swap the keys is still in the
server; this message might be totally vestigal.  
The message almost certainly doesn't have anything to do with parity errors.

If this option isn't implemented and you'd like to implement it, look at
the original IBM code.  All this option does is lie about key caps; it 
physically swaps the control and caps lock entries in the modifier and 
key caps tables.   It's probably 10 or 15 lines of code.

Good luck!
   Erik
   (erik at sgi.com)

Disclaimer:   Needless to say, I didn't work for SGI when all of this was going
     on.   I bear full responsibility for the contents of this article.