On Feb 7, 2008, at 12:30 AM, J R wrote:
I disagree with you. it is IMO an OPEN issue it should return a RC
that indicates that open was successful but as a sequential dataset.
A return code after a successful OPEN is already too late.
CLOSE will wipe out the directory.
Yes/NO it depends on *IF* the *DEFAULT* is to write over the
directory if its *not* going to then a RC is OK if it is going to
write over the directory then I agree with you. Again it could be a
parmlib option at IPL time. Since no sysgens are really done anymore.
It would be interesting to listen to a discussion of the IBM people
and how they would look at it and we might get more of an insight as
to how much code and what options there might be.
Ed
As Ted says, if you say you want to open the dataset as PS,
that's what it does.
Going back to the '60s design, they could have included
some sort of open option that would indicate that you were
intentionally overriding the DSORG, without which OPEN
would fail.
I'm curious -- does anyone remember how OPEN behaved
if you opened an ISAM dataset with DSORG=PS?
Date: Wed, 6 Feb 2008 23:53:15 -0600
From: [EMAIL PROTECTED]
Subject: Re: VSAM Surprise
To: [email protected]
On Feb 6, 2008, at 3:31 PM, Ted MacNEIL wrote:
I think it violates the principle of least astonishment. Writing
to a PDS as a PS dataset is almost certainly done by accident. It
would be nice it IDCAMS would just say "You didn't want to do
that.".
It's not an IDCAMS issue.
It's a documented (flaw) feature of JCL.
The first place searched for file attributes is the executing
programme.
If it thinks the file is sequential, it is treated as such.
Ted:
I disagree with you. it is IMO an OPEN issue it should return a RC
that indicates that open was successful but as a sequential dataset.
Although it might break a lot of code its a close call IMO. I think
it revolves around OPEN and a feedback code (which is not presented
to the end user) so IDCAMS is not really at fault its OPEN that has
the flaw, although I can see it from both sides. I can see where IBM
would say BAD and walk away from it, or say its fixed in rel 9999.
Its a no when situation as IBM coded open the way they did and who
knows how much code would need to be fixed? Now if someone had caught
this back in the 60's maybe.. but not this late in the game. On the
other hand they could make IDCAMS (if this is coded in the JCL create
a member called tempname. This might also break programs, but
probably less than the first.
Ed
_________________________________________________________________
Helping your favorite cause is as easy as instant messaging. You
IM, we give.
http://im.live.com/Messenger/IM/Home/?source=text_hotmail_join
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
