Open-source Languages & Tools for z/OS

The simplest python program print(‘hello’) fails with 9.

JCL is:
//BPX EXEC PGM=BPXBATSL,
// PARM=‘PGM /C/Rocket/python36/bin/python /a/private_folder/test.py’
//*
//STDOUT DD SYSOUT=*
//STDERR DD SYSOUT=*
//*
//STDENV DD *
PATH=$PATH:$HOME:/C/Rocket/python36/bin
PYTHON_HOME=/C/Rocket/python36
PYTHON_PATH=/C/Rocket/python36/lib/python3.6
LIBPATH=$LIBPATH:$PYTHON_HOME/lib
LANG=C
LC_ALL=en_US
TERM=xterm
_BPX_SHAREAS=YES
_BPX_BATCH_SPAWN=YES
_BPXK_AUTOCVT=“ON”
//*

STDERR contains some garbage:
…?..%…%…`…?>……?../…>?..?.>…?…?…%…>…

If I change BPXBATSL to BPXBATCH with sh instead of pgm in PARM, then job is successful, but I’m going to use DDNs in the future scripts. Is there any idea why the code fails and how to see normal letters?

Python 3.6.1

Vladamir
I am facing exactly the same issue. Any solutions you have found so far?

Thank you
Nagaraj

Nagaraj,
I saw that only with Python 3.6. With Python 2.7 I see normal error message. Since version 3.6 can’t work with opening data sets using DD cards i.e. f = open(’//DD:INDD’, ‘r’) doesn’t work, I use only Python 2.7 where there is no garbage.

Hi Vladimir
Thank you very much for your quick response. My JCL looks like as below. I am simply trying to print(“Hello World”) in the program hello.py which is /u/sys1156/temp directory. I am getting a RC of 9 with garbage printed in STDERR. I am not reading any instream data in any of the DD cards in my JCL as you can see below. The stdparm and envvars files I used are also shown below. I am still getting a RC of 9. Any ideas?
//SYS1156B JOB (1TPF0160),‘BPXBATCH’,CLASS=A,MSGCLASS=X,
// NOTIFY=&SYSUID
//*
//STEP01 EXEC PGM=BPXBATSL
//*
//STDPARM DD PATH=’/u/sys1156/temp/stdparm’,
// PATHOPTS=ORDONLY
//STDERR DD PATH=’/u/sys1156/temp/pgmout’,
// PATHOPTS=(OWRONLY,OCREAT,OTRUNC),PATHMODE=SIRWXU
//STDOUT DD SYSOUT=*
//SYSPRINT DD SYSOUT=*
//SYSOUT DD SYSOUT=*
//*
//STDENV DD PATH=’/u/sys1156/temp/envvars’,
// PATHOPTS=ORDONLY
//*---------------------------------------------------------------------
BROWSE /u/sys1156/temp/envvars

PY36=/usr/lpp/python/py36/python-2017-04-12-py36/python36
LIBPATH=/usr/lpp/python/py36/python-2017-04-12-py36/python36
PYTHON_HOME=/usr/lpp/python/py36/python-2017-04-12-py36/python36
_BPXK_AUTOCVT=“ON”
_BPX_SHAREAS=MUST
_BPX_BATCH_SPAWN=YES
******************************** Bottom of Data *****************
/u/sys1156/temp/stdparm

pgm /usr/lpp/python/py36/python-2017-04-12-py36/python36/bin/python
/u/sys1156/temp/hello.py
******************************** Bottom of Data *******************

1. Try to execute it in omvs to see how it works there
2. Then try to use EXEC PGM=BPXBATCH,PARM=‘SH program-name’ to see how that way works
3. Then try to use EXEC PGM=BPXBATCH,PARM=‘PGM program-name’ to see how it works. Here you need to provide your ENV also.

If all 1 and 2 ways work fine and 3 fails, then obviously problem in your env. Something might be missing. You need to put everything from your .profile to be sure that you’ve got all the needed. Keep in mind that in .profile you can have symbolic variables like $PATH. In EVN data set you need to provide paths as they are.

Hi Vladimir
Thank you very much for your quick response again. The problem was that I was browsing the STDERR file, instead when I viewed/edited it using ISPF 3.17 (Z/OS UNIX Directory List) option, what was being shown as garbage started showing up as readable text.
The problem in my program was that I did not create the file with appropriate encoding. I followed the below advice given in this forum, but as a response to a different question, after which it all worked:
<<Advice I followed - begin>>
I would recommend you to delete existing file, create the file again (touch test1.py), tag it (chtag -tc ISO8859-1 test1.py). After that add a content into the file and run it.
Thanks,
Tatyana
<<Advice I followed -end >>