Rocket Mainframe DevOps

The download area shows: git-2.3.5-src_b013.170517.tar but no accompanying binary tar archive. Is this intended? Thanks, Manfred

Status: DeliveredDeliveredi would like to suggest that Serena deliver a member in zmf control file (CNTL) for the define and initialization of cmnvpool dataset that is used by HLL exits.  Right now we have to CUT and PASTE the JCL required for define from delivered manuals, it would be much easier if JCL was delivered with product.

HI I am working on a operating system Z/OS v2r2 Http server is installed and is correct on my lpar. /etc/webap1/bin/apachectl -v Server version: IBM_HTTP_Server/9.0.0.0-PI54808 (Unix) Server built: Jan 20 2016 17:19:40 Php is the tool of Rocket and is installed in : /PRODUITS/ROCK/php EUID=0 /PRODUITS/ROCK/php/ Type Filename _,Dir ,. _,Dir ,… _,Dir ,bin _,Syml ,CHANGES.ZOS _,File ,fichier.php544.tar _,Dir ,include _,Dir ,lib _,Dir ,man _,File ,README _,Syml ,README.ZOS _,Dir ,share _,Syml ,VERSION.ZOS /PRODUITS/ROCK/php/bin/php -v PHP 5.4.4 (cli) (built: Nov 21 2016 03:25:49) Copyright © 1997-2012 The PHP Group Zend Engine v2.4.0, Copyright © 1998-2012 Zend Technologies I made the export in the .profile of the user of the stc. (start server Apache) export PATH=$PATH:/PRODUITS/ROCK/php/bin export MANPATH=$MANPATH:/PRODUITS/ROCK/php/man/ When I want to see my page html, the command (php) are not interpreted. Html page : Standard HTML Page with PHP HelloWorld <?PHP pr

We have the 4.2.53 version of bash and are having problems with BPXBATCH when using the exec command. bash -version GNU bash, version 4.2.53(2)-release (i370-ibm-openedition) Copyright © 2011 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later http://gnu.org/licenses/gpl.html If the last line of my .profile contains an exec for the bash shell When running a batch TMP with BPXBATCH it produces no output. (ie appears to run but just returns the READY prompt). This is what I ran //ADDIT EXEC PGM=IKJEFT01,REGION=0M,TIME=1440 //SYSTSPRT DD SYSOUT=* //STDOUT DD SYSOUT=* //STDERR DD SYSOUT=* //SYSTSIN DD * BPXBATCH SH ls -la /* Same results for PGM=BPXBATCH,PARM=(‘SH ls -la’) This is what is in my .profile exec /sysn/rs/bin/bash

I have installed most of the Rocket ported software, just for my own use (it’s good to be a sysprog). I am reading up on TCL and using it in learning mode on my Linux desktop. In the book, and on Linux, the TCL scripting command is “tclsh”. In your port, in /tcl-8.5-11/bin there is the command “tclsh8.5” with a symlink to it of “tcl”. I think it would be more consistent with what I’ve seen elsewhere if there were also an symlink of “tclsh”. Of course, in the “UNIX” arena, I am a GNU/Linux aficionado, so I prefer my z/OS UNIX to looks a bit like what I’m used to in Linux.

Now available for download on the Rocket Community portal, please find a new release of cURL for z/OS. This port of v7.52.1 was implemented with z/OS enhanced ASCII support. This means it will work with both ASCII and EBCDIC files but be sure your files are tagged. Don’t forget to read the README.ZOS file.

Hi, I’m running bash script with the bash 4.3 implementation and notice a, to me, strange behaviour. The script file does not have any tag and I get one error when trying to execute it. After the try, the script file has been tagged with ISO8859-1. When trying again, I get another error since the script is no encoded in ASCII. I then tag (chtag) the script file with EBCDIC (IBM-1047) and then the script runs perfectly well. So far so good, but I have some scripts that are delivered to me in a RO filesystem. When running one of those, I get the first error and of course the script file will not be tagged after the try. The older version of bash, 4.2 does not have this behaviour. It runs my scripts perfectly well without any tagging whatsoever. Would it be possible to direct 4.3.46 to behave as 4.2? I have the following relevant environment variables set. _BPXK_AUTOCVT=ON export _CEE_RUNOPTS="FILETAG(AUTOCVT,AUTOTAG) POSIX(ON)" export _TAG_REDIR_ERR=txt export _TAG_REDIR_IN=txt export _T

Just downloaded then ftp’d to z/OS -rw-rw-rw- 1 RXT025 OMVSDIS 296615 May 2 13:53 unzip-6.0_b0006.161123.tar.gz installed via /bin/cat /zusr/home/RXT025/gz/unzip-6.0_b0006.161123.tar.gz | gunzip -c | /bin/tar -C /zusr/home/RXT025/unzip -xUXovf - result running is cd /zusr/home/RXT025/unzip/bin RXT025@SYSS:bin, Tue May 02, 02:16 PM $ unzip -l í>!Ñø?à øÊÑ%Âñ>Ã?!ñ&(/Ñ>È/Ñ>ÁÀÂäëøÑÁ%ÁÊëÁ>ÀÂÍÅÊÁø?ÊÈËÍËÑ>ÅÇÈÈøÏÏÏÑ>Ã?:Ñø?ÊÅ:ÑøÂÍÅÇÈ_%ËÁÁêá à(áÃ?ÊÀÁÈ/Ñ%ËíË/ÅÁÍ>:Ñø$!)$?øÈË$?ÀÑÃÑÁÊË))ÃÑ%Á$:Ñø)$%ÑËÈ)$ÌÌ%ÑËÈ)$ÀÁÌÀÑÊ)àÁÃ/Í%È/ÄÈÑ?>ÑËÈ?ÁÌÈÊ/ÄÈÃÑ%ÁËÑ>%ÑËÈÁÌÄÁøÈÈÇ?ËÁÑ>Ì%ÑËÈÈ?ÁÌÀÑÊÃÑ%Á$:Ñø)/ÂÁ/ÏÑ%ÀÄ/ÊÀ!!Ñøñ>Ã?_?ÀÁÍ>:Ñø!Ã?ÊÍË/ÅÁøÁÌÈÊ/ÄÈÃÑ%ÁËÈ?øÑøÁ>?_ÁËË/ÅÁË%%ÑËÈÃÑ%ÁËËÇ?ÊÈÃ?Ê_/ÈÃÃÊÁËÇÁ>ÁÌÑËÈÑ>ÅÃÑ%ÁËÄÊÁ/ÈÁ>?>ÁÈÈÁËÈÄ?_øÊÁËËÁÀ/ÊÄÇÑÎÁÀ/È/ÍÍøÀ/ÈÁÃÑ%ÁËÄÊÁ/ÈÁÑÃ>ÁÄÁËË/Ê:ÀÑËø%//ÊÄÇÑÎÁÄ?__Á>È?>%Î%ÑËÈÎÁÊÂ?ËÁ%ËÇ?ÏÎÁÊËÑ?>Ñ>Ã?èÈÑ_ÁËÈ/_ø/ÊÄÇÑÎÁÈ?%/ÈÁËÈÌÁÌÄ%ÍÀÁÃÑ%ÁËÈÇ/ÈÃ?%%?ÏÑ>Ì%ÑËÈÀÁÌÈÊ/ÄÈÃÑ%ÁËÑ>È?ÁÌÀÑÊ_?ÀÑÃÑÁÊË>>ÁÎÁÊ?ÎÁÊÏÊÑÈÁÁÌÑËÈ

If you attempt to clone or checkout a very large git repository (one with a great deal of history), you may get this message: bash-2.03$ git checkout --track -b v0.12-zos-ascii origin/v0.12-zos-ascii fatal: Out of memory? mmap failed: EDC5124I Too many open files. (errno2=0x07360344) bash-2.03$ The problem is that git, by default, attempts to mmap the entire pack file (in the .git/objects/pack directory) into memory, and in this case that’s a very large file (about 235MB): bash-2.03$ ls -l .git/objects/pack total 494592 -r--r--r-- 1 jcallen ussuser 7578432 Jan 26 11:23 pack-13c65cae0f429287bcff36df4d582d30125aa766.idx -r--r--r-- 1 jcallen ussuser 245499367 Jan 26 11:23 pack-13c65cae0f429287bcff36df4d582d30125aa766.pack bash-2.03$ The z/OS port is 31-bit program, so that represents a sizable chunk of the entire 2GB address space. Fortunately, git provides some knobs to control memory use. A stackoverflow item provided the necessary clues. I used these commands to crank down th

git log gives error: cannot run less: EDC5129I No such file or directory. as a first line. Perhaps you should add some documentation in the included pdf document about how to set up git log to use more because there is no less out of the box in z/OS UNIX. There are either environment variables PAGER or GIT_PAGER or git specific core.pager config option. – Manfred

Tired of having to enter your information in order to download z/OS open source? Well we did not like the old system either and we finally have implemented a new one. Now, anyone, whether a Rocket customer or not, can download any z/OS open source product from Rocket’s Community Portal. If you do not have a login to the Community Portal, you still need to enter your information, but you will only need to do this once. To login to the portal or to request an account go to this page: https://my.rocketsoftware.com/RocketCommunity/RCLogin Once you have logged in, you will see a Downloads button at the top of the page. Select the category “z/OpenSource” on the left side of the page and the list of products will appear on the right side of the page. Click the link to the left of the product you want and then download the version you want.

I took a look at the perl 5.24 installation which is Tool: perl Version: 5.24.0 Build Number: 002 and found that its directory structure has some oddities. There are duplicate directories lib/perl5/5.24.0/os390/ lib/perl5/5.24.0/5.24.0/os390/ Most of the files are duplicates but not all as arithmetic shows: $ find lib/perl5/5.24.0/5.24.0/ -type f | wc -l 1628 $ find lib/perl5/5.24.0/ -type f | wc -l 3151 – Manfred

Release notes for Git 2.3.5 for zOS.pdf says on page 2: perl 5.24 Note: This version of perl is not yet appropriate for general use. Is this still true? – Manfred

We have a software product that makes extensive use of Perl 5.8.8 – yes, I know, that’s an 11-year-old version. We’ve tried a couple of times to update to more recent versions of Perl, but with all the Perl modules that our code depends on plus our own custom modules it rapidly turned into a compatibility nightmare, and both times we gave up. Now we’d like to port our product to z/OS, so we’re looking for a Perl ~5.8.x version for z/OS. I believe IBM used to provide 5.8.7 (which should also work for us) as part of their Ported Tools program http://www-03.ibm.com/systems/z/os/zos/features/unix/ported/perl/ , but they withdrew support for it in 2013, and I don’t know where to find a copy (nor what the legal status of using such a copy for commercial purposes would now be). Does Rocket provide a Perl version of around that vintage (I don’t see one on the website), or does anyone here know where if might find or how to obtain such a thing (possibly a copy of the Ported Tools one that IBM u

Subject: Creating a custom report Description: I am looking to create a report from CM ZMF V8.1.0, with the following information: Package names (for a given time range) List of SRC components in each package Work release # Creator Lines changed in each SRC/ Total lines of code in that SRC There are some reports close, but nothing with SOURCE LINES CHANGED/TOTAL SOURCE LINES. I can obtain this info by using package lists, and further use of 'VC' and compares, but would like a report or XML that I could run more readily in batch or online from TSO. Does anyone have a similar report or advice?#ChangeManZMF

Announcing availability of Python 3.6.1 beta 1 for z/OS. To download, go to Rocket Corporate web site download page: http://www.rocketsoftware.com/zos-open-source/python?ver=3.6.1 Multiprocessing still not implemented. Expect beta 2 in a few weeks with Anaconda support. Expect beta 3 in a couple months with multiprocessing support.

The beta 2 release of Python 2.7.12 for z/OS is no longer available. Instead there is now a beta 3 release of python 2.7.13 and a beta 1 release of python 3.6.1 (see other post). Python 2.7.13 To unpack this distribution: gzip -d -c python-2017-04-12-py27.tar.gz | tar xoUXf - Then follow instructions in PYTHON_README.ZOS. This release fixes the following: · SciPy and many other math, science, graphics and other packages are now included · File tags are now heeded when reading text files

When using -z option (zEDC), is it possible to unzip in other plataforma, like Windows?

I installed CURL 7.42.1 on our mainframe z/OS USS file system and then tried to run a curl command for the first time using batch BPXBATCH. curl -v http://www.yahoo.com/ It got this message “IDN support not present, can’t parse Unicode domains”. What does this message mean?

TL;DR: Is there a way to properly encode received bytes using EBCDIC ? Hello, while working with sockets I noticed that python always expects and EBCDIC string. I have currently been able to get a Python client-server configuration (Linux client, z/OS server) to communicate by forcing the client to use ‘cp500’ encoding, but I haven’t been able to do the same server-side (decoding Latin-1 to EBCDIC). This is what I tried up until now: >>> import codecs, sys >>> test = ‘\\x48\\x54\\x54\\x50’ >>> test.encode(‘cp500’) UnicodeDecodeError: ‘ascii’ codec can’t decode byte 0x48 in position 0: ordinal not in range(128) I also tried by setting the default encoding to ‘cp500’ with: >>> reload(sys) >>> sys.setdefaultencoding(‘cp500’) >>> test.decode(‘Latin1’) u’\\x48\\x54\\x54&’ Still, without success. Any hints?

The following code gets an RC32. I can't figure out why, all the substitutions appear to be correct when I run a trace. Can anyone show me the error of my ways or another way to get the auditReturnCode in this V8.1.2 HLLX exit? If function = 'FREZ0101' then Do say " " say "------------------------------------------------------" say "Package Freeze HLLX call " say "------------------------------------------------------" say "Function : "function" "Date()" "Time() proceed = "YES" dataChanged = 'YES' /*****************************

Having installed Tool: make Version: 4.1 Build Number: 0002 now running make I get rubbish characters ./make /,�������+?��/����ˀ����������/>��>?�/,���%���?�>�����?��# Actually the output is ascii and when translated is: make: *** No targets specified and no makefile found. Stop. So it looks to me that make 4.1 is unusable. Or did I miss anything? – Manfred

I tried to use curl to call a REST API that returns a JSON payload. The response was not converted to EBCDIC. I tried with an XML response which was the same. I’m assuming that curl uses the MIME type to figure out what to convert. So any text/* MIME types get correctly converted. I would suggest application/json should also be converted which by the standard is UTF8 encoded. application/xml is different as XML can be in any encoding but I would take a punt and also convert it to EBCDIC.

PHP 5.4.4 Build Date Nov 21 2016 03:23:06 running on z/OS 2.2 under Apache V9. It seems that PHP does not take into account the she-bang on the first line when reporting the line number in error, therefore it is always one less than its actual location in the script. The following example code contains a syntax error on line 7: 000001 #!/path/to/php/php-cgi 000002 <?php 000003 000004 $v = “A value”; 000005 000006 // The following line contains a syntax error 000007 $a = arra(); 000008 000009 ?> When run in the browser the following message is displayed: Fatal error: Call to undefined function arra() in /my/cgi on line 6 Notice that the reported line number is “6” when the actual error occurs on source line “7”. If the same script is run using PHP 5.1.2 (previously supplied by IBM) then this message is displayed: Fatal error: HPH01209 Call to undefined function arra() in /my/cgi on line 7 Now the correct line number is displayed. I guess it’s debatable as to whether or not the she

Using Rocket PHP 5.4.4 Build Date Nov 21 2016 03:23:06 on a z/OS 2.2 system under Apache V9. The fopen() function can successfully process a local file but experiences an ABEND0C4 when attempting to process a URL (e.g. http://…). The following code is an example: $url=“http://www.rocketsoftware.com”; **$handle = fopen($url,“r”); ** **if($handle) { ** ** fclose($handle); ** ** echo “Successful fopen/fclose”; ** **} ** else ** ** echo “ERROR: Could not fopen file”; When this code is executed in a browser a 500 Internal Server Error page is displayed. The following message appears on SYSLOG: IEF450I PUBLIC STEP1 - ABEND=S0C4 U0000 REASON=00000004 The following messages appear in the Apache HTTP server error log: AH01215: CEE3204S The system detected a protection exception (System Completion Code=0C4).: AH01215: From entry point _php_stream_write_mvs at compile unit offset +000000C2 at entry offset +000000C2 at address: AH01215: 299AD60A.: End of script output before head