I checked in the missing pages; Now the only errors are due to
fpsum and the missing "sys read".
The cold=1 build does seem to write the file system. There's still a
problem however. I notice that that after the initial file system is
all written, the code does an RF read with a large positive word count
(this is bad, since the WC should be negative). I'll be it's the
exec.
I think the exec of "init" is causing this, and the same thing happens
when I run a cold=0 load after building the file system. But who knows.
So, we're closer. Many thanks to James for offering his OCR version.
I'll put it up on my web site later today and send out an email,
along with the diff files.
-brad