I did a comparison of s2's /bin/sh and the sh.s we have (notes in
the notes file about this). There were a few changes, but they were
pretty similar. I added a patch in the patches directory that reflects
the changes and I verified that if you build with the patch you get
an identical binary (modulo the padding bytes that the assembler
fills in for ".." and ".even"). It's currently marked as an optional
patch and the sh.s build isn't being installed over /bin/sh in the
rf0.dsk image.
I think the sh.s listing itself might be defective in two ways:
- There's an rts missing at the end of the "error" function
- It doesn't check for the "*" character when looking for globbing.
The 1e man page does list "*" as one of the globbing characters,
so its likely that these three instructions (1 rts, 2 for comparison)
somehow got dropped from the commented jun72 listing. I'm torn between
adding them to the original "pages" sources as obvious mistakes or
just leaving it out and using a patch to fix it. Right now we don't
have a patch that adds just these two fixes. I guess if we did that and
made it a mandatory patch, we could use the sh we build from sh.s as
/bin/sh in the images.
I look forward to seeing what shell sources we might get from the s1
fragments recovery.
Tim Newsham
http://www.thenewsh.com/~newsham/
if you want to contribute, but don't have e.g. arcane knowledge of
PDP 11 assembly and such -- then let me suggest that it would be
interesting to find out more about these people listed in the 1973
"Study of Unix" documents (http://www.bitsavers.org/pdf/bellLabs/unix/)
that formed the basis of this reconstruction effort.
For starters, who was this "T. R. Bashkow" who called the
meeting? Some googling last week indicates to me that he has
an engineering award named after him, and that he does
not have a wikipedia entry.
B. A. Tague's name is prominent too, although I personally
do not recognize it. And similarly for the other memo
addressees.
Consider that any of these people might just happen to still
have source code listings, magnetic/DEC-tapes, paper tape, or even just
historical anecdotes to share, but perhaps no one ever asked
them.
Or their heirs, for any who have passed on.
In any case, it's getting kind of late in history; this may
be the last chance to track down even information about these
people who participated in this historic meeting, let alone
find them and ask them about ancient media.
P.S. The above thoughts apply to all other historic systems, of course,
not just those Unix-related.
I managed to find a RESISTORS-related historic document last week,
for instance. (Princeton-area early hobbyist computer group.)
Doug Merritt
--
Professional Wild-eyed Visionary Member, Crusaders for a Better Tomorrow
I just committed some changes that reorganize the build process a bit.
The Readme contains information about the change. Basically now you
just do a "make" at the top level. If you have a checked out tree
you might have to move your current "build" aside when you do an
update. Please let me know if anyone notices brokenness.
Warren, can you look over the Readme?
Tim Newsham
http://www.thenewsh.com/~newsham/
An early section of the jun72 document has some code for
"modifications to UNIX to accomodate the T4002A graphic console."
Is anyone familiar with this? If we could emulate this device
easily, it might be fun to enter this patch as an optional patch
to the system.
Tim Newsham
http://www.thenewsh.com/~newsham/
Ok, I had some initial success with a quick and dirty hack.
- apply the attached patch to your simh tree and rebuild pdp11
and copy it into your tools directory.
- maps the TTIX device into the DC11 space
- forces the CSR to return the carry-detect bit as always on
- edit build/init.s and uncomment some or all of the ttys
build it and place it on your drive:
$ vi build/init.s
$ tools/as build/init.s
$ cp b.out build/root/etc/init
$ tools/mkfs -p fs/Readme build/root rf0.dsk rf
- copy your simh.cfg to simh2.cfg and add these lines before
the "go" line:
set ttix en
set ttix lines=8
att ttix 5555
- run "./simh2.cfg"
at this point you can telnet to port 5555 and enjoy a multi user
unix system. This lacks a lot of polish and needs a bit of work
yet, but its quick and dirty.
You'll have to revert to the old init if you want to use the normal
simh.cfg since the kernel will panic when opening the other ttys
if the DC11 hardware isnt attached.
Tim Newsham
http://www.thenewsh.com/~newsham/
> We need to recreate enough of the boot chain to make self-support possible.
> James Markevitch I think was working on this; James, can you give us a
> status report?
It's up and running. It requires the KE11 fix that Brad made (so that
switch registers work). I tested against that and can boot without
pre-loading core with anything other than the 32W bootstrap. It pulls
the bootstrap (bos) and operating system from the high portion of the
RF11 image.
I'll upload this this evening.
James Markevitch
Wednesday I got V1 running on my Mac. Great fun!
The level of sophistication impressed people here.
Do you folks think we can move the sources for
the kernel into it and have it be self supporting?
Hi
I have a few xt computers and would like to put some old unix OS on,
what unix would you guys recommend for it ? most of the unix that i have
tried to search for are for machines that are older than xt.
Regards,
Angus
I've just written a Readme to describe how to build the kernel and filesystems,
and to run 1st Edition UNIX. Can a few people who haven't actually used the
system please have a go and give me feedback on the Readme.
To fetch the repository:
svn checkout http://unix-jun72.googlecode.com/svn/trunk/ unix72
Then go into the new unix72/ and read the Readme.
Thanks,
Warren