Re: RFC: Alternate patch to have true new-style rc.d scripts in ports (without touching localpkg)

I guess that by having enough people bring up their concerns, I am
hoping that such a change will be reconsidered...

Without sounding too much like a broken record - separating packages
from the base into a different partition is an asset I would much
rather not loose.  I believe many people have made a coherent enough
argument that I wholehearedly agree with, but I'd just also like to
add something I haven't seen (perhaps missed).

I have a setup where an old slow NIS/NFS server provides a combination
of /usr/local, /home, and swap for a bunch of even slower machines.  The
beauty of this setup is that as long as the software installed in /usr/local
is built for the lowest common denominator processor (all are obviously
x86) everything just works.  And it works well.  The scripts that run
from /usr/local/etc/rc.d run exactly as expected, on all machines, and
I am happy to only change them in one place to have the whole farm
"just work".  In all this, there is but one kludge - the dhcpd startup
script just exits if run on any host other than the one hardcoded in
the script.

I would imagine that this setup is not unique to me - but it relies on
having the startup scripts in /usr/local/etc/rc.d;  I gladly accept
criticism on this (e.g. this setup is dumb, here's a better way to do
it) or explanations why startup scritps for packages in /etc will not
break the setup above.  But please don't offer me to also start NFS
mounting /etc/local; /var/db/pkg is already annoying enough.

Mike