> hmm, that is strange. I wonder what is triggering this and making ksh
> die.

I have a moderate-to-strong suspicion that this, ironically, will involve
/bin/ksh on Solaris being really ksh, where we probably have some packages
tested against NetBSD's /bin/ksh, which isn't, and developers testing against
a ksh that isn't are likely to end up with ksh scripts that aren't.

Now that ksh is CPL'd it should be easier to nip that problem in the bud.
I'm looking at updating shells/ast-ksh as a step in that direction.

-Chap

btw, what does the current Solaris ksh respond to
$ print ${.sh.version}
?