[TUHS] Determining what was on a tape back in the day

Will Senn will.senn at gmail.com
Sun Nov 19 07:26:55 AEST 2017

On 11/18/17 12:57 PM, Clem Cole wrote:
> A quick look, and I think it's an stp (super TP) tape -- stp is from 
> the Harvard distribution.   This would make sense, because that was 
> the standard back before tar.
> As Ron pointed out, tp (which Ken designed for DECTapes originally) 
> puts the index at the head of the tape (tar and later cpis threaded 
> the index inline).   But it means its a fixed size and there were some 
> other issues (tp may have originally been in assembler IIRC).   On 
> DECtape, tp worked pretty well/was pretty cool because you could 
> update a block, less so on 9-track which when you re-wrote block N, 
> you lost all blocks afterwards.  Also, I don't remember why now 
> [probably the limits off the directory], but it was typically in those 
> days to take all the files in a directory, turn them into a foo.a (ar 
> format) archive.  So the stp image was a bunch of files: dir1/mumble.a 
>  dir2/grumble.a dir3/bumble.a ...
> You then needed to unarchive the files within each directory.   Also, 
> remember ar(1) when through some changes in format between 4-7th 
> editions as the compiler and linker matured.  So watch out on that 
> front too...
> Anyway, v6 tp probably will read it, but if you poke around the TUHS 
> and bitkeeper archives for the original Harvard distribution, stp.c 
> should exist.
I'll look around. v6 tp was able to read the tape:
set tc en
att tc0 unix6.dat
# chdir /usr/6
# tp t0

the directories don't get created on extract, but that's typical on v6.


GPG Fingerprint: 68F4 B3BD 1730 555A 4462  7D45 3EAA 5B6D A982 BAAF

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20171118/25c533d8/attachment.html>

More information about the TUHS mailing list