Yes, I actually, personally encountered this message.
Yes, it substituted the keyword which was, from memory, <GIMUNZIP>.
I was kind of mystified because there in the job was
//SYSIN DD *
<GIMUNZIP>
etc.
The problem turned out to be one of my other favorite design -- umm, things I
would have done differently had I written MVS -- the generated SYSIN statement.
Preceding the above in the job was some misplaced non-JCL record and the
interpreter or converter or whoever generated a gratuitous SYSIN DD statement,
and that is the SYSIN dataset that the message was complaining about.
> But I do wish that when SMP/E reports a syntax error it would print a ruler
> with a pointer indicating the exact point of error
OT, but may I brag? The product I DID design and implement reports errors as
shown in this example.
CZA0082S Error detected while processing maximum message length ('40000') of
'MAXMSGLEN'
CZA0084I Near column 31 of record 5 of Parameter File
CORRELOG.CZAGENT.CNTL(CZAPARMS)
CZA0105I V
CZA0106I SERVER 10.2.1.118 MAXMSGLEN(40000)
CZA0023S Value must be numeric and between 512 and 32767
Charles
-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf
Of Paul Gilmartin
Sent: Wednesday, September 07, 2016 7:45 AM
To: [email protected]
Subject: Re: FTP on z/OS sucks
On Wed, 7 Sep 2016 06:42:19 -0700, Charles Mills wrote:
>An example from yesterday. Not sure if this message is owned by USS or
>not but it sure fits the pattern you describe. Pasted in its entirety
>without editing.
>
>GIM37150S
>
GIM is SMP/E's prefix.
> ...
>GIM37150S
>THE keyword KEYWORD IS MISSING OR INVALID.
>Explanation
>
>keyword
>missing or invalid keyword
>
Did you actually experience this message yesterday?
What did it substitute for "keyword" If no substitution, I's call it a bug
rather than a design deficiency.
But I do wish that when SMP/E reports a syntax error it would print a ruler
with a pointer indicating the exact point of error, rather than a column number.
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
