MB> KK> But what happens if both, keyboard and mouse, are completely
  > KK> 'cutted off' by another misbehaving thread in such a way,
  > KK> even WatchCat's device driver doesn't receive a message from
  > KK> the keyboard?

MB>In theory, this should never happen.  The keyboard
  >generates a hardware interrupt, IRQ1.  The system should
  >service that immediately unless interrupts are disabled or
  >some higher priority interrupt is stuck pending.  When that
  >happens, the device driver will have control in an
  >interrupt context.  If the driver has also hooked a timer
  >or gets control in some other way, then a PM lockup will
  >not affect the driver.

When using Borland's 1.0 in the debug mode, I have found that 'watching'
variables, say in 'main', and during the switching between the
app-window and the ide, the system will occasionally hang solid.  Even
on the second WatchCat attempt, IT will also hang.
___
 X OLX 1.53 X Psychoceramics: The study of crackpots.

--- Maximus/2 3.00