[rescue] Configuring SLIP, both (Unix) server and client ends (Dell Sys V rel 4 Unix)

[Mouse]

>> [...] I somehow made it thru the whole .COM thing without ever using
>> or configuring SLIP or CSLIP.
> You didn't miss much.  SL/IP was a pain.

It could be a pain.  Occasionally it was (and is - see below) a
lifesaver.

In particular, I have run into situations where negotiation and such
are impediments, not assets.  If I just have two machines on the bench
and want them to talk, SLIP is a *lot* less pain to set up than PPP:
connect the wire, slattach and ifconfig each end, and boom! working.
PPP, at least in the implementations I have access to, requires
designating one end as a server, setting up a bunch of configuration
files, and a good deal of diagnostic work to figure out why it's not
working.

Most recently, I was - less than a year ago - trying to get a machine
networked when the OS I had didn't understand either the Ethernet or
the USB hardware.  But the machine had six serial ports.  It would boot
off a USB thumb drive, but the kernel didn't understand the USB
hardware, so it was unusable after boot.

I took half an afternoon to hack in support for automatically bringing
up SLIP at boot time and it Just Worked.  Based on past experience
trying to set up PPP, it would have taken at least twice that long,
probably more - and the SLIP version included writing the in-kernel
code to make it all happen at boot time.

> Or even writing the basics of a file system level FUSE driver seems
> easier than trying to get anything modern to talk SL/IP

Then you need better "modern" systems.

If you have anything like NetBSD's or Linux's tun driver in your kernel
(and IMO there is no excuse for anything claiming to be modern to lack
it), you can even do the rest of it in userland.  SLIP encapsulation
borders on trivial; even if you use my hacks to support non-IPv4
packets (which you may want to do if you have tap but not tun), it's
only marginally less trivial.

> Once PPP was usable, it took over in a flash.

For client-to-server connectivity, perhaps.  I still find it more of a
liability than an asset for "two machines on the bench, just want 'em
to talk" connectivity.

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		mouse at rodents-montreal.org
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B