Help on a CTIX machine

[was-John McMillan]

In article <2592 at ncsuvx.ncsu.edu> mikes at ncsuvx.ncsu.edu.UUCP (Michael Steele) writes:
>  I am new to this newsgroup so pardon any ignorance regarding the topic.
>  As a part of the South Hall Computer Theme Program at NC State University
>  in Raleigh, NC, we aquired a Convergent Tech Mini-frame plus a few months
>  ago.				^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^- o my...

	As a matter of fact, the Mini-frame was, at one point,
	not much more than an overweight UNIX(rg)-pc: sort of
	the price one paid for a coupla extra connectors.  If
	you are NOT running SVR3, this is probably the most
	appropriate hole in which to post questions!

	Yes, a bit of history -- as distorted by my weak
	recollections -- of things:

		The 3B1 & Mini-frame software had the same --
	or very common -- base, many years ago.  As I recall,
	there were numerous 'ifdefs'.  For various contractual
	reasons -- proprietary issues, mostly -- CT's product-
	line kernel staff were separated from the AT&T-product
	kernel staff and an unknown [to me] amount of divergence
	ensued.

>        Below are several questions I have regarding the system, any help
>  you can provide will be GREATLY appreciated (either by posting to the net
>  or via email).  We have a mini-frame plus with v3.10net of SYS V.
						  ^^^^^^^^^^^^^^^^^^
	Since I haven't the faintest idea of what "v3.10net" is,
	I can't address it directly.  The UNIX-pc is derived from
	CTIX: the UNIX-pc was NOT BASED on SVR3 -- although much
	of the SVID [System V Interface Definition] is common.
	
>  1) Is there an FTP site for pc sources (ie an archive) or software that is
>  configured for the 7300 3B1 miniframe, etc (I've heard they are all binary
>  compatable).  What has been a pain for us is converting BSD programs to
>  Sys V, I'm sure others are doing the same.  Has someone archived these
>  conversions?

	It is quite possible that CTIX remains binary-compatible with
	the UNIX-pc.  Divergence [from Convergence ?-) ] in the Shared
	library is always a threat.  Or, subtle variations in the
	system calls could have crept in:

		AT&T spent considerable effort maximizing SVID
	compliance for the UNIX-pc.  CT may not have incorporated
	those in their product.  Conversely: they may be MORE in
	compliance by virtue of running a full SVR3.  (If you
	have /usr/include/sys/region.h you're SVR3.)
	AT&T did attempt to contract with CT to deploy the CT SVR3
	"product" (I never really saw it, and don't know if was
	ever really released for the Mini-frame) on the 3B1, but
	negotiations (or work) collapsed as strains accumulated
	-- typical when a LARGE and small corporation work together.
	[No comment.]

		Toward the end of our period of depending on CT to
	do the 3B1 kernel work, we were told, a number of times,
	"we've already fixed that in our version" -- indicating
	they had isolated AT&T contract work from their own BUG
	FIXES!  (Save electrons: you can't add anything meaningful
	to OUR comments.)  This may indicate further divergences,
	even if you aren't running SVR3.

	While you CAN avoid shared libraries, they are generally a REAL
	blessing in disk/swap savings.
		
>  3) Has anyone seen a PD korn shell for the CTIX machines?  

	Has anyone seen a P[ublic] D[omain] korn shell for anything?

>						  	Is it true 
>  that AT&T distributes UNIX free to universities...I assume this
>  information is VERY old, or do they still do it.  If so, how would one go
>  about upgrading their software and can we get kernal source?

	'Think there was always a nominal fee: in the early days
	it was little more than a tape-charge.

	Try phoning:	(800) 828-UNIX	-- and post your results.

>  4) I've seen several postings about gcc running on AT&T machines.

	Try loading a STRIPped 'gcc' from someone's UNIX-pc.  If it
	can run, can't you bootstrap to a full, long-names ["flexnames"]
	development environment?  (The UNIX-pc moved to "flexnames"
	around release 3.0 [or was that 3.5?].)

John McMillan	-- att!mtunb!jcm	-- muttering for himself, ONLY