> :-( Any chance of maybe increasing this "out of hours" say 4am to 11am
> your time :-> I guess thats there has to be a time when no-one's logged
> on doing the hacking ?

umm, "guess again" -- just about the only time is during dumps,
when i kick people off...  we've actually got enough people
in enough places (europe, australia, etc.) that it never slows down
*that* much.

> 'cause I still haven't managed to sup the new currect across since that
> big update. I run over a 14.4 ppp line, froma site in London, with 19
> hops from me to you...

"that last big update"?  i.e. the rcs id one?  well, when the new tar
files are made this saturday morning, you'll be able to grab them
from either agate or lamp (via ftp!), and that'll get you started...

> > With 12 max-connections, 40 disk irq's per second, and
> > over 100 irq/sec ethernet. Squish, squish, squish.
> 
> ergh ! What is lamp ? Hardware wise that is... My 486/33/1742/32Mb/S3
> *feels* like my IPX / LX at work - and seems as quick to compile as
> a SPARCstation 10/30 (not a 2 processor 41, like our development server
> tho' :-)

that's on a 486DX-50, 1742, 3 *FAST* disks, 16M ram, nobody using
console.

> BTW - I agree that development takes priority, but if it's this difficult
> to get -current, is there now chance you could find cpu-time on another
> box and make that the sup server ?

umm, i'd like to do something like this -- actually, i'd
rather put all ftp and mail and sup traffic on one machine, change its
name to 'sun-lamp', and use *another* machine for development.

the problems are:
	(1) finding the time to set it up
	(2) finding the hardware...  lamp has come out of my own
		pocket, and its sister machine (pain, which has
		a small disk, and is used for testing kernels, etc.)
		did too.  i've got another machine that i'd like
		to make such a server, but the hard disk controller
		currently doesn't work with NetBSD, etc...



chris

------------------------------------------------------------------------------