Problem with JCL

[magnus]

Hi.

I'm having trouble with a COPY, my JCL looks like this:

//COPY EXEC PGM=COZBATCH
//STEPLIB DD DSN=COZ.LOADLIB,DISP=SHR
//STDIN DD *
fromdsn -s IBM-1143 -t ISO8859-1 -l crlf -k //ZSHR.TCPIP.TESTFIL \
> /u/user/testfil

If i do this via a command in OMVS it works, but my JCL don't.

I suspect that it's something wrong with the \ since if I remove the transalation stuff and do a plain copy on one line it works in the JCL.

And the errortext:

CoZBatchÝN¨: Copyright (C) 2005-2009 Dovetailed Technologies LLC. All rights re
CoZBatchÝN¨: version 1.10.1 2011-11-07
CoZBatchÝI¨: executing progname=login-shell="-/bin/sh"
fromdsn(\)ÝE¨: DatasetHandler: Error in fopen(//'\', rb,type=record,noseek,recfm=*) - EDC5047I An invalid file name was
specified as a function parameter. (errno2=0xC00B0287)
CoZBatchÝI¨: returning rc=exitcode=0

Br.Magnus

[magnus]

Well, as I thought it was the \.

It's always interesting to run on a different codepage, of course my contunation sign is É and not \.

Problem solved....

[dovetail]

Good news. FWIW, line termination processing is done by the z/OS Unix shell (or whatever program you ran in COZBATCH), and not COZBATCH.

Agreed, z/OS Unix is kindof a mess when it comes to locales & encodings. Somethings are locale/encoding aware, some things are not.

[magnus]

Good news. FWIW, line termination processing is done by the z/OS Unix shell (or whatever program you ran in COZBATCH), and not COZBATCH.

Agreed, z/OS Unix is kindof a mess when it comes to locales & encodings. Somethings are locale/encoding aware, some things are not.

Yes of course line termination is handled in z/OS UNIX, that was what came to my mind, I suddenly remebered that I have had this problem in shell scripts before, that's how I found the É sign to be "my" \

And YES z/OS UNIX is a MESS in encoding, it's a mix of EBCDIC and ASCII and on top I'm not running US codepage...

Br.Magnus