HP3000-L Archives

October 1996, Week 5

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:
Reply To:
Date:
Mon, 28 Oct 1996 10:13:07 -0800
Content-Type:
text/plain
Parts/Attachments:
text/plain (36 lines)
Cortlandt Wilson <[log in to unmask]> wrote:

> [snip]
> Q3.  As I understand it a DBXUNDO applies to all DB calls across multiple
> databases. Do I have this right?

I don't know, but there is something else you may have to worry about.

You mention (later) that this will be deployed under MPE/iX 4.0 and 5.0.
I know there was a situation a few years ago where DBXUNDO would be
required in certain circumstances, and these conditions CHANGED from 4.0 to
5.0. (See pages 6-16 and 6-17 of the 5.0 Communicator.) It happens that the
"new" way is the way it should have been in the first place (IMHO), but be
aware that your app may run differently under 5.0 than under 4.0.

An example of this is if you try to update a record by first calling
DBUPDATE. If that fails because a search or sort item changed, then you can
call DBDELETE / DBPUT. (The intent here is that you wouldn't have to modify
your application if a chain was added or deleted.)

In a DBXBEGIN / DBXEND block under MPE/iX 4.0, this strategy FAILS because
a modify intrinsic (DBUPDATE here) failed, so the ENTIRE TRANSACTION has to
be cancelled. Under 5.0, though, the strategy works.

> Q4.   Any "gotcha's" one should know about?

There may be a problem when using DBXBEGIN / DBXEND with NetBase.
I *think* a previous client traced a system abort to this under MPE/iX 4.0;
I don't recall which version of NetBase. You may want to check with Quest
about this if it may affect you.

--Glenn Cole
  Software al dente, Inc.
  [log in to unmask]  <-- client site
  [log in to unmask]      <-- always

ATOM RSS1 RSS2