In a recent note, "SPITZ, HOBART (CONTRACTOR)" said:
> Date: Mon, 6 Aug 2007 13:41:52 -0400
>
> I can't find it now, and I may have misread it, but I was responding to
> what I thought was a suggestion to use SPEC 1-* C2X 1.
>
I was unaware of SPECS C2X. Mea culpa.
> Melinda's approach has the advantages of being well documented,
> extensively tested, and already performance tuned. She used a uuencode
> technique.
>
> See her paper entitled "Streamlining Your Pipelines."
>
I looked at it. AFAICT, it fails to distinguish between
linefeeds added by the stage:
| 'spec 1-* 1 x0A next |', /* Append line-feed to each. */ |
.. and linefeeds initially present as characters in the EBCDIC text.
> I think the "digest" stage is the way to go. And because it's common
> standards you can actually compute the digest on both platforms and
> compare.
>
.. but I don't have DIGEST (yet), nor can I count on all of our
customers having it (yet). Does DIGEST provide different sums
for inadvertently concatenated records? If not, "PACK | DIGEST"
is the way to go (if one can count on DIGEST's being available).
-- gil
--
StorageTek
INFORMATION made POWERFUL
