🚀 go-pugleaf

My question about disk space got a good response (thanks everyone), so let's
try something even more arcane. 

  I've used CallAfter and CallEvery before, but never for an accurately-timed
delay. I've just been testing this, and it seems that they both run about 10%
slow. Specifically, if I ask to be called after 100 ticks (1 second) I actually
get called after something between 109 and 110 ticks. The average number is
quite stable, but doesn't appear to be an integer. I'm measuring the slippage
relative to the monotonic time, which does seem to be accurate. I find this
odd, as I had assumed that they were clocked off the same interrupt. Does
anyone know what the problem is? Can I rely on it being 10% slow? What is there
that happens at slightly more than 9Hz??? I wasn't trying it on a "virgin
machine"; is it possible that a rogue module is messing it up?

  BTW, this is another area where the PRM is deficient; it doesn't say what the
entry and exit conditions for the called routine are, other than "treat it as
an interrupt routine". Do I have to preserve registers? And is the processor in
IRQ mode? It would be nice to have an OS_CallBackAfter, since whenever I use
CallAfter I always seem to need a CallBack as well. Not in RISC OS 3 by any
chance? :-)

e----><----p | Stephen Burke           | Internet: burke@vxdesy.desy.de
 H   H   1   | Gruppe FH1T (Liverpool) | DECnet:   vxdesy::burke (13313::burke)
 H   H  11   | DESY, Notkestrasse 85   | BITNET:   BURKE@DESYVAX or SB2@UKACRL
 HHHHH   1   | 2000 Hamburg 52         | JANET:    sb2@uk.ac.rl.ib
 H   H   1   | Germany                 | Phone:    + 49 40 8998 2282
 H   H 11111 | HERA, the world's largest electron microscope!