I thought I would pass along something I just discovered about an old
tar tape from 1987.
One of our scientists brought me a stack of old 9-track tapes (circa
1987-1990) to read this afternoon. All but one were ANSI-labelled
"VAX COPY" tapes. The one that was not said it was a Unix tar tape.
I have a tape scanning program that told me it had a single file on
the tape with 10240 byte blocks, which matches what you should see on
a tar tape. I read it on to our MicroVAX as a foreign tape with
RECORDSIZE=BLOCKSIZE=10240. But, when I tried to use tar to read it
(VMSTAR on our Alpha, tar on my Mac, tar on a Solaris SPARC
workstation), it kept saying it was not a tar archive. I opened the
file on my Mac using TextWrangler. There were lines that looked like
data, and text that looked like it could be a file name, but I didn't
recognize any of the text. I figured at least it was an uncompressed/
unencrypted archive of some sort (thank goodness). I thought it might
be the old Unix dump tape format, but I could not find out what that
format was like (I didn't spend much time Google searching). I looked
once more at the ASCII in the archive, and mentally byte-swapped some
of the strings that looked like they should have been file names.
BINGO! What looked like gibberish became words. I used dd conv=swab
on my Mac to byte-swap the file and, lo and behold, tar -tf worked.
About the time the tape was written, we had a pair of VAX-11/750's
networked to a VAX-11/785. One of the VAX-11/750's ran BSD Unix. I
don't know whether there was a device name option to select byte-
swapping (in those days, at least, Unix used different device names
for the same tape drive to select among the capabilities of the
device, such as to select the density) or if BSD always byte-swapped
when writing tapes on a VAX (I doubt it -- this is the first time I
have ever encountered a byte-swapped 9-track tape). Whatever the
reason, I thought it might be useful for this group to know that such
tapes exist. Thus, if you get a real tape of a tape image that
purports to be in tar format, but is not, don't give up -- it could be
byte-swapped!
Larry Baker
US Geological Survey
650-329-5608
[email protected]
_______________________________________________
Simh mailing list
[email protected]
http://mailman.trailing-edge.com/mailman/listinfo/simh
