On 11/10/97 at 5:51 PM -0600, you wrote:

> You'll have to create the partition under Mac OS, as native disk labels
> are not supported (and won't be until after the 1.3 release, at least).

Hm. Okay. (Man, I'm *never* going to have a respectable uptime. :) I
haven't broken a month yet. I've never had a crash, excepting times I've
booted with unstable kernels to test 'em. Oh well.) (But then, I guess I
couldn't hot swap the new drive in, even if I *was* able to use disklabel
under NetBSD. Ah, well.)

Random note:

A guy I'm interviewing with runs HP-UX on some obscenely large system. He
was telling me how some RAM went bad one night on the system, and HP-UX
detected the fact, automatically sectioned it out, and rebooted. Supposedly
the next version of HP-UX will be able to handle the problem without even
having to reboot. Can we support cool stuff like that? I'm guessing that
HP-UX would die screaming if the memory containing the kernel itself went
bad, and I'm not sure how our internals work, but if my understanding is
correct, we know to look for paged out stuff if we try to access some
memory and get an error, right? Maybe this feat could be accomplished by
taking advantage of that same sort of error, or whatever sort of error we'd
get with bad RAM.

Just curious. Maybe someday I'll feel competent to try to get something
like that to work. I'd imagine running with an open case with an evil
intent in my eye and my hands firmly grasping a SIMM would be an adequate
test. (I think I can scrounge up a 386 to use as a testbed.)

--
Mason Loring Bliss...mason@acheron.middleboro.ma.us...www.webtrek.com/mason
"In the drowsy dark cave of the mind dreams build their nest with fragments
 dropped from day's caravan."--Rabindranath Tagore...awake ? sleep : dream;