HP3000-L Archives

October 1999, Week 2

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
"Danny A. van Delft" <[log in to unmask]>
Reply To:
Danny A. van Delft
Date:
Thu, 14 Oct 1999 22:20:13 GMT
Content-Type:
text/plain
Parts/Attachments:
text/plain (44 lines)
On Thu, 14 Oct 1999 11:08:24 -0400, Lars Appel
<[log in to unmask]> wrote:

>Did I say "can of worms"? For the grins... My C.60.00 machine knows
>about -A in help tar, but looks more like yours regarding man tar...
>

Well, I'm always in for a good laugh but when I ran into this one it
took at least an hour before I could laugh again. I (gratefully) had
been using tar since MPE5.0 to distribute various archives to various
sites, both my own (well, you know ...) as well as external, of
various os's (MPE/iX, HP-UX, MS-Windows  (well, not exactly an
operating system ...), DEC, Sparc, and even MS-DOS (well, not exactly
an ...). Suddenly I got reports from these sites that the tar files
produced unexpected results or failed to untar. Luckily I'd had the
foresight (and the disc space) to copy the 5.0 SYS account to an
OLDSYS account prior to the update, bit of an overkill but y'never
know... I therefore could temporarily create a link from HPBIN to
point to the "old" TAR before I found out where the problem originated
and got the -A solution from HPCRC.

And then that pesky problem of not always correctly untarring certain
files. I had hoped with the 5.5 version to completly get rid of MOVER
and ARCXL and move to a more "standard" tool, but that's not the case
(yet, I hope). For example, I'm using Allbase/SQL with preprocessed
modules that need to get transferred to other systems. These modules
are in files with 250W fixed binary records. It is not amusing to
transfer these modules in a tar archive to the target system,
untarring them, seemingly without a problem, and just finding out that
they're corrupt when you try to install them into the DB. The "funny"
part is that, while they look ok, (right number of records, right
record length etc.) only by doing a cmp with the original module
(which you don't have on that system and might have been purged on the
originating system, they're safe in the archive aren't they?) you find
out they are different.

Sorry for the above, in no way is it directed at you. Just wanted to
"share".

All in all, I am still very pleased with POSIX on MPE. It makes most
things more easy to do. Still some work to be done though in that
area, but, what else is new...
Danny A. van Delft   [log in to unmask]

ATOM RSS1 RSS2