bjorn@sysadm.UUCP (Administration Login) (04/28/88)
I have seen many postings discussing the UNIX-PC software. Unless some people out in net-land are better of then I am, you are running Unix V release 2. The release 3.x mentioned in the spec's for the machine are the version of of Unix V.2 for that machine. I am making this posting in order to avoid other people getting caught as I have been. When I ordered my machine I chose the UNIX-PC in order to get a vanilla ATT Unix V.3, as all specs I ever have seen talks about release 3.x. If anyone has infor- mation which contradicts the above, please post. Bjorn Satdeva uunet!sysadm!bjorn
stevens@hsi.UUCP (Richard Stevens) (04/29/88)
In article <135@sysadm.UUCP>, bjorn@sysadm.UUCP (Administration Login) writes: > you are running Unix V release 2. > The 3b1 UNIX is truly a hybrid of System V, Release 2.0 (enhanced version) and System V, Release 3.0. It contains the demand paging that came out with the enhanced version of System 5.2. *But* it also contains mandatory (yes, mandatory) record locking, and shared libraries, which didn't officially appear until System 5.3. However, it doesn't contain the other goodies that appeared with 5.3: RFS, Streams, TLI, TPI. It looks like they took the 3b1 release somewhere between 5.2 and 5.3. One note about the record locking: the fcntl(2) and lockf(3) man pages imply that the record locking is either advisory or mandatory, and both reference you to the chmod(2) man page for details. However, nothing on the chmod(2) man page mentions it. If you have the 5.3 manuals, you'll find that what chmod should have mentioned is that *if* you set the sgid bit on and the group-execute bit off for a given file, then the record locking for that file is mandatory, not advisory. Under 5.3 the ls(1) command recognizes this and prints 'l' to indicate that mandatory record locking is enabled for the file, but the 3b1 ls doesn't do this. All the more evidence that they took the 3b1 version somewhere between 5.2 and 5.3 and never got back to it and never got all the pieces together (i.e., documentation) for what they did ship. Richard Stevens Health Systems International, New Haven, CT { uunet | ihnp4 } ! hsi ! stevens
alex@umbc3.UMD.EDU (Alex S. Crain) (04/29/88)
In article <135@sysadm.UUCP> bjorn@sysadm.UUCP (Administration Login) writes: >I have seen many postings discussing the UNIX-PC software. Unless >some people out in net-land are better of then I am, you are >running Unix V release 2. Well, sort of. the Unix on a 3b1 is not complete sysVr2, but has things not found in sysVr2 as well, liked shared libraries. I like to call it "A subset of SysVr2 with SysVr3 extensions" :-) -- :alex. nerwin!alex@umbc3.umd.edu alex@umbc3.umd.edu
friedl@vsi.UUCP (Stephen J. Friedl) (04/29/88)
In article <135@sysadm.UUCP>, bjorn@sysadm.UUCP (Administration Login) writes: > I have seen many postings discussing the UNIX-PC software. Unless > some people out in net-land are better off then I am, you are > running Unix V release 2. The release 3.x mentioned in the spec's > for the machine are the version of Unix V.2 for that machine. I've been amused by the talk of Sys V Rel 3 in this group as well. The UNIX-PC software is almost certainly based on System V Release 0. Our original manuals for the 7300 mention SVR0, and lots of other signs abound: /bin/sh has no shell functions, the miserable curses, the old archive formats, and the dates found in various system files. My primary experience is with UNIX-PC Software version 3.0, and it looks *very* much like SVR0. Velease 3.5, which I am only slightly familiar, seems to be a lot better (terminfo, etc.). Nevertheless, I find it highly improbably that AT&T (or Convergent) built 3.5 from fresh SVR2 source; they likely just added the user-level utilities to Release 0. The only real SVR3-ism I see in the UNIX-PC is the shared library facility. Rebuttals of this are encouraged. -- Steve Friedl V-Systems, Inc. "I do everything in software, even DMA" friedl@vsi.com {backbones}!vsi.com!friedl attmail!vsi!friedl
jon@jonlab.UUCP (Jon LaBadie) (05/01/88)
In article <953@hsi.UUCP>, stevens@hsi.UUCP (Richard Stevens) writes: > The 3b1 UNIX is truly a hybrid of System V, Release 2.0 (enhanced version) > and System V, Release 3.0. It contains the demand paging that came out > with the enhanced version of System 5.2. *But* it also contains > mandatory (yes, mandatory) record locking, and shared libraries, > which didn't officially appear until System 5.3. However, it doesn't > contain the other goodies that appeared with 5.3: RFS, Streams, TLI, TPI. > It looks like they took the 3b1 release somewhere between 5.2 and 5.3. I must strongly disagree! True, paging and shared libraries (and record locking) are in both systems. But both bats and birds fly. This does not make them related. The facilities mentioned were implemented by AT&T (for SVR3) and by CT (for release 3.0 of the 7300) in totally independent environments. Their solutions are not compatible, are implemented differently, and are used differently. I do not know of a single item, new to AT&T's SVR3 that has been incorporated into the UNIX-PC. Had the PC succeeded, the plan was to make release 4 of the UNIX-PC software compatible with SVR3. HIGHLY unlikely that will happen now. Release 3.5 was SUPPOSED to be SVR2 compatible! However, there are SOOOOOO many items in 3.5 that were not upgraded to SVR2, that I would drop Richard's categorization of the UNIX-PC's software to between 5.0 and 5.2, not 5.2 and 5.2 Examples of NON-5.2 utilities include: cat ls crontab at vi mailx uucp and on and on ... Jon LaBadie {ihnp4, ulysses, princeton}!jonlab!jon
darren@bacchus (Darren Friedlein) (05/02/88)
I have no understanding whatsoever of 5.0 or 5.2, but I know that mailx
and at aren't included. What do you do without ls, cat and uucp, tho?
-Darren
____
/ \
| Rt 4, Box 416, Durham, NC 27703
_____|_____ Darren G. Friedlein data (bacchus) : 919/596-7746
/ | \ voice : 919/596-9492
( | )
\____/ __/ {mcnc|icus|ethos|gladys|bakerst}!bacchus!darren
wtm@neoucom.UUCP (Bill Mayhew) (05/02/88)
I agree with Steve Friedl that much of the so-called Unix on the Unix PC is Sys V R0. The release 3.5 and 3.51 seem to incorporate quite a bit of stuff from Sys V R1. There is also a lot of BSD-like stuff in there too -- take for instance the presence of more in place of pg (as the first that comes to mind). I notice that AT&T is careful to say in the manuals that the UNIX PC software passes SYSVID for 5.2; they avoid saying it *is* 5.2. It is annouying that the dates on most of the system files are at Jan 1, 1970 so that the incremental back-up won't back them up. Tsk tsk. Also helps hide how old some of the files are too.. --Bill
clb) (05/03/88)
In article <430@jonlab.UUCP>, jon@jonlab.UUCP (Jon LaBadie) writes: > In article <953@hsi.UUCP>, stevens@hsi.UUCP (Richard Stevens) writes: > > The 3b1 UNIX is truly a hybrid of System V, Release 2.0 (enhanced version) > > and System V, Release 3.0. .... > I must strongly disagree! > True, paging and shared libraries (and record locking) are in both systems. > ... > ... Had the PC succeeded, > the plan was to make release 4 of the UNIX-PC software compatible with > ... > Had the PC succeeded? Mine succeeds just fine. In my view there is an enormous gulf between what AT&T sees as a success and true success. The measure of computers should not be entrusted to so many greedy bean counters. The 7300 works much better, much more reliably and much faster that the pdp-11 (version 7) or the Altos xenix boxes that I have worked with in the past. You have to be a nit-picker to find things that don't work, whereas many other systems have so many bugs that they crawl all over. And any one who tries to do useful work with an MS-DOS machine knows what a mess that is. Just because the company that (doesn't) supports the machine can't succeed for their own folly is no reason to blame the machine, it's design or it's function. Money is all they understand... not excellence or utility or value. Just the almighty buck. As an engineer and programmer, I reserve the right to decide for myself. Charles Brunow clb@loci.UUCP
dca@kesmai.COM (David C. Albrecht) (05/04/88)
> > ... Had the PC succeeded, > > the plan was to make release 4 of the UNIX-PC software compatible with > > ... > > > > Had the PC succeeded? Mine succeeds just fine. In my view there > is an enormous gulf between what AT&T sees as a success and true > success. The measure of computers should not be entrusted to so > many greedy bean counters. > Working just fine and succeeding are two different issues and I dare say greed has little to do with it. In the UNIX PC AT&T tried to promote the rather absurd concept that business users should pay more for an O/S they didn't understand on single brand hardware with limited applications. Business users are interested in applications not O/Ss and justly the UNIX PC landed with a resounding thud. Even as a low end workstation it was simply too pricey and with its only fair performance was generally greeted with ho-hum (remember now that the original 7300 1/2M 20M machine went for 5K+). AT&T simply doesn't have the marketing savy and production efficiency to be a price leader where the machine may have made a dent. So now, AT&T has reduced the price of the 7300/3b1 machines to the point that they are selling for less than they cost to produce to get rid of the inventory before they are so mouldy as to be worth nothing to nobody. Certainly, this is a great deal for those of us who are willing to accept its limitations in exchange for a great price. Certainly, it is a quite repectable unix box. A success? The amount of money AT&T lost on the UNIX PC was probably phenomenal (to us mortals anyway). If your idea of a success is a product and pricing that people like but the company loses its shirt on I hope you stay out of corporate America (they have enough problems). David Albrecht