I suspect you are correct.  On a SPARC 2 with Weitek upgrade (80MHz),
running with "startx -- -ar1 1" produces an acceptable repeat.  Turning
off XKB with "startx -- -kb" works too.  I'll probably just turn off XKB
since I'm not sure what XKB buys you right now, with few programs
supporting it.  A quick scan through the XKB extension shows that key
repeat is mentioned in about five different source files so it'd be tricky
to trace where the actual incompatibility lies. 

---Jake

On Tue, 28 May 1996, James E. Bernard wrote:
> 
> >One minor problem:  I built X11R6.1 straight from the X Consortium
> >distribution, and there is no keyboard repeat in the X server, even though
> >xset says that it is turned on!  This is very annoying.  X11R6.1 includes
> 
>   I have the same problem on a couple of sparc 1's.  Except, the key repeat
> isn't actually missing, it's just that the delay in initiating autorepeat
> is excessively long (approaching a minute on one system, about 5-6 sec on
> another).  The -ar1 arg to the server does have an effect, but, at least on
> the one system on which I've tweaked it repeatedly, even setting it to
> 1 ms leaves the delay unacceptably large at about 1 sec.  I suspected a
> simple scaling problem in the units used for the value, but found none in
> a brief perusal of the code.  The apparent difference in behavior on two
> identical machines leads me to suspect something weird about the mechanism
> used to time the delay, but I won't have time to look into it further for
> a few weeks.
> 
>   One other thing I've noticed is that the beep function sometimes makes
> exceedingly short, almost non-existent, beeps.  Other times it seems more
> or less normal.
> 
> --Jim Bernard
>