>Unfortunately, AFAIK mh-e isn't supported under MicroEmacs... :')

Nor is anything in real e-lisp.

>Seriously, having to fall back on MicroEmacs makes NetBSD/pmax not
>seem like a real operating system, which it is.   Making it work on
>the pmax is effort that should be useful on other NetBSD platforms
>eventually - the Elf32 undump problem isn't a pmax-specific problem.

True. But emacs _does_ ``work'', in the sense that emacs-20.2 from the
FSF configures and builds and runs cleanly on NetBSD/pmax

It just(!) doesn't unexec.  Yes, there's good grounds to call that
`broken', but it works well enought to have discouraged me from
getting it `right'.  It would be a fun project for a dedicated
hacker-wannabe.


Frankly, I think Ted is right: the ELF unexec in Emacs code is a total
mess. There are separate versions for close to every supported RISC
CPU/OS combination.  Or maybe it's just mips and alpha. 
(Mh, I guess that means it doesn't work on Linux/mips either?)