Greetings all,

Well, I've got a -current system up on my trusty 4MB portable, and
the good news is that it seems a lot more robust than the -current
from just before Christmas. Kudos and thanks.

[I eventually ftp'ed the binaries from ftp.cursci.com:/NetBSD-stuff,
untar-ed them, brought up a 0.9 kernel with the -current
binaries/libs, and made a -current kernel.  (Thanks, Adrian).  I'm
currently making the libs.  Thanks to Mike Long for the information
about building make, tsort and gas with the OLD include files, and
then using the new binaries and includes from there - I didn't try it,
but it sounds right.  Thanks also to the person who told me I
shouldn't be going from 0.9 to -current if I'd already done that once,
and thanks to the person who sent me Marc Wandschneider's
how-to-upgrade instructions].

Observations:

Charles Hannum's fix to swap_pager.c is in the kernel I'm running at
the moment.  Having said that, there are times when the compilation
just seems to hang, doing nothing - no disc access, zilch.  So I
suspended the make and put it in the background.  This seems to work
for a while, and then it hangs again.  However, when I press return, I
get the usual csh prompt back, and if I type fg, it tells me no such
process, but the compile suddenly comes to life again, and off we go. 
Is there a missing call to wake up a thread somewhere? Tickled by some
of the job control trickery?

I like the new messages from the wd driver.

The VM changes look good - this is running on 4MB memory, remember,
and, while it may not be fast, it does work.

Alistair

------------------------------------------------------------------------------