> A new binary snapshot of 1.2-A is available on ftp.netbsd.org.

Which reminds me that I should have reported some problems that I had with
the previous snapshot (not that they're repeated in the new one, but while
I'm thinking about it):

Don't use the install.sh script if you're installing over an existing system.

I was running 1.0 and decided to install the April 12th snapshot over it.
install.sh assumes that the /, /var and /usr are all mounted at the same time.
Since my customary install method is to boot a SunOS CD-ROM into the miniroot
and install from there, that's what I did.  I burned a CD-R disc with the
snapshot and mounted it from the SunOS miniroot.  I then proceeded to let fly
with "install.sh" (using the SunOS "gtar" binary to unbundle) and all Hell
broke loose.  My /usr and /var partitions were newfs'd under NetBSD, so when
install.sh started extracting stuff into those partitions (under a SunOS
miniroot, remember), it totally mangled the root inodes of each of them.

Thus, quickly it filled up the /usr partition because for every directory in
my existing /usr, "gtar" created a *new* directory *with the same name* because
of the root directory mangling.  Yup, two "bin"s, two "include"s, and so on.
Do Not Do This At Home Kids.  It took quite a bit of work to clear up all of
the shrapnel.  I ended up only extracting the root stuff (I'd originally
newfs'd the root partition under SunOS when installing 1.0, so I wasn't worried
about mangling things), then did a "boot -bs" into single-user mode with only
"/" mounted to do the rest of the extractions in a NetBSD environment.

I also recall there not being a "telnet" in the April 12th snapshot.  I had my
old 1.0 telnet binary lying around, so it wasn't critical, but I couldn't
find one.

Also, the extracting had some weird hiccups.  Some programs had changed from
being symlinks to being real binaries (or maybe it was the other way around?)
and it wouldn't overwrite the existing stuff with the new version of the
differing type.  I seem to recall problems with "more" because of this (since
Matt changed it to being "less"-based between 1.0 and 1.1B).  There were a
couple of other cases of this as well.  I had to go to my Pentium 120 system
running a then-current to see what was really what.

There were problems with include files being missing (under /usr/include/g++
in particular), but I think this has since been addressed.

Anyway, there were a couple of other things, but I've forgotten them now.
I figured the best way to cure all the problems that I didn't notice right
off the bat was to bring it up to -current, so I did that a week ago.

Slight aside:

BTW, is anybody else in NetBSD/SPARC land using a SPARCstation 2 (or IPX) with
a Weitek power-Up chip?  I had mine go belly-up last week.  Strangest thing.

I brought the system down and powered it off to add an Exabyte 8200 to the bus
(many strange problems with trying to use this under -current, btw), and ever
since that time it's refused to work.  After a short period of time, the system
would just freeze up solid.  Thanks to "options BLINK" I knew it was dead  :-)

Anyway, I'd power-cycle it and it would come back up, then freeze again.  At
first I thought it was my new kernel; rebooted with the old one and it froze
up anyway.  Gradually the freezeups happened earlier and earlier.  I decided
to get NetBSD out of the loop and booted a Solaris 2.5 CD-ROM.  Froze up again.
Finally I L1-A'd it as it was coming up, left it at the PROM prompt, and the
phone rang.  Talked for 5 minutes, hung up, then went to type "boot" and it
was frozen again!  Would be curious to know if anyone's heard of a similar
failure mode with the Weiteks.  Mine was working fine for 9 months and we
have another SS2 similarly equipped (main NIS and mail server as well - doh!)
which is still running fine.  Most peculiar.

	- Greg