Path: utzoo!attcan!lsuc!eci386!woods
From: woods@eci386.uucp (Greg A. Woods)
Newsgroups: comp.sys.att
Subject: Re: remote execution etc. on StarLAN
Summary: 3b1 r-utils should be good for any SysVr3.x TLI
Message-ID: <1990Oct23.172540.4590@eci386.uucp>
Date: 23 Oct 90 17:25:40 GMT
References: <3750@rossignol.Princeton.EDU> <1990Oct18.113228.7342@cbnews.att.com>
Reply-To: woods@eci386.UUCP (Greg A. Woods)
Distribution: na
Organization: Elegant Communications Inc.
Lines: 31

In article <1990Oct18.113228.7342@cbnews.att.com> mvadh@cbnews.att.com (andrew.d.hay) writes:
> it was a port of BSD rlogin et. al. to *STARLAN-1* for the 3b1.
> you'll probably find it on osu-cis.
> 
> while you might make it work on a 3b2/starlan-1, i don't believe such
> was ever offered for the 386.  on the other hand, you don't need it
> for starlan-10 -- isn't it just ethernet over twisted pair?  can't you
> just run the BSD stuff as is?

After a cursory glance at these tools, it appears a small amount of
clean-up will make the port work on any TLI implementation.  I.e. the
port has changed the r-utils to use the t_* networking routines from
SysVr3.x, i.e. libnsl_s.a, as described in the SysV Network Programmers
Guide.  I.e. as long as you have a homogenous network using these
tools, it doesn't matter what the transport layer, or the physical
layer for that matter, is.

[The clean-up probably means better integration with the listener,
i.e. the way HDB-UUCP is.]

As for "just running the BSD stuff as-is", you can't, unless you have
a full BSD socket library or emulation library.  This is possible on
many TCP/IP implementations for SysVr3.x, though such implementations
usually come with ports of the r-utils.

There's a big difference between TCP/IP and Ethernet.
-- 
						Greg A. Woods

woods@{eci386,gate,robohack,ontmoh,tmsoft}.UUCP
+1-416-443-1734 [h]  +1-416-595-5425 [w]    VE3-TCP	Toronto, Ontario CANADA