Product Forums

Hi,I have found a changes in behaviour with the sql disconnect function between native COBOL (using ODBC) and .NET COBOL (using ADO) and was wondering whether these are deliberate or notScenario: I have a primary SQL connection ‘MAIN’ which is used for the majority of processing. However I have two other connections ‘ADMIN’ and ‘ALTERNATE’ which are used to perform elevated functions (ADMIN) or functions which I want disconnected from main (ALTERNATE)In native COBOL when I disconnected from one of these secondary connections it automatically reverted to the MAIN connection however in .NET COBOL this is not the case and I am finding I need to issue a ‘set connection’ .Also in native COBOL when the connection is closed it physically kills the connection to SQL - i.e. the SPID is no longer active. In NET COBOL whilst the connection is logically closed I see that the SPID still exists in SQL.  Are these changes deliberate? Peter 

Let’s address  the AI-lephant in the roomHi everyone, Welcome to our monthly Rocket ACUCOBOL newsletter. This edition tackles the biggest question of the month (or maybe even the year): How is GenAI changing ACUCOBOL modernization?Code Translations vs. Modernization:What’s the difference This week, COBOL made headlines again as the debate around AI‑generated COBOL resurfaced. The market reaction and the conversations online point to one thing: change is no longer coming. It’s here.Though after more than 5,000 successful COBOL and ACUCOBOL modernization projects across all the industries, we’ve learned one lesson that matters now more than ever:Code translation is not the same as software modernization. And confusing the two can put enterprises at real risk. Code Translation Is Useful. But Language Isn’t the ProblemMany AI companies promise rapid “COBOL modernization,” but what they’re really talking about is syntax translation. And while COBOL is a straightforward language (ask any COB

You can find the most recent ETX hotfix in KB Article 000088551.Please note that the download is available to active maintenance customers with Rocket Community login.

Hi ,  since January (2026) we started the preparation to implement Changeman ZMF 8.3.1.  Investigation was ongoing. By accident , when checking some info on your website mid February , we discovered that since February 10 th (2026)  version 8.3.2 is available.   Since we discovered this by accident , we were surprised.Is there a way we can be informed when a new version is available ?  (We are already subscribed to newsletters)Or is no communication done on this topic ?

Due to some security issues, the D3 installers' ability to automatically download dependency packages is no longer available. This makes it necessary to load dependencies manually before running D3_setup. Here are the instructions:First, disable D3_setup's attempt to load dependencies:mkdir /usr/lib/pick  ( if /usr/lib/pick doesn't already exist )cat /dev/null > /usr/lib/pick/d3depends  ( if the depends file doesn't already exist )==================================Manually loading dependencies on Linux versions 7 and lower:D3 Linux versions 9.2 and lower:yum install dialog.i686yum install glibc-develyum install ncurses-develyum install libgcc.i686yum install net-toolsyum install gccyum install pam-develyum install pam-devel.i686yum install glibc-develyum install glibc-devel.i686yum install libuuidyum install openssl098eyum install hal-libsyum install libidnD3 Linux versions 10.1 through 10.3.2:yum install dialogyum install dialog.x86_64yum install gccyum install libcurl-develyum ins

I am moving from Bluezone to Rocket TE and do not have macros in Rocket.  Is there a means to convert my bluezone VBA macros to a script that can be used with Rocket TE?

Our Cobol apps run natively on a Linux host with just local logging files, and I was wondering if there is any documentation on emitting (Opentelemetry) traces to an external collector (via network).If not natively supported, maybe people have some other ideas how to bridge the gap? Maybe building some integration via “Consolidated Tracing Facility” (CTF)?Thanks for any idea!Tung

Uniface and EDIFACTAuthor: i.sharp@pcisystems.co.uk (Iain Sharp) Has anyone used the EDIFACT standards to create/consume an EDI solution?  I'd like to determine if I need a third party app to convert to something useful, or if I can write something which will work directly in Uniface (or as a COM/DLL plugin).  If anyone has done similar, could you briefly say how difficult a task it was?  Regards,  Iain

I have a customer who is using Shimms Universe database. Can anyone guide me to someone who can help build APIs and extract data for this customer? Any help or guidance would be greatly appreciated.

Our team uses an Excel VBA macro to automate actions through Rocket Terminal Emulator (BlueZone). It was created on machines and is used by users that have Excel 32‑bit and the macro runs without any issues. New team members, however, are on Excel 64‑bit, and they are all getting the following error: Have anyone one else had similar errors and what was your solution?

Announcing the latest update for ChangeMan ZMF 8.3 Patch Update 2 (8.3.2)Rocket Software is pleased to announce a new release of ChangeMan ZMF, delivering a significant set of enhancements focused on modernization, resilience, usability, and deeper lifecycle automation. This release contains more than 50 enhancements, including over 20 driven directly by customer-submitted ideas through our Ideas portal.Below are some of the key highlight themes from this release:Enhanced Change Package Visibility & ControlsThis release introduces several improvements that streamline Change Package management and strengthen lifecycle governance. Enhancements include clearer visibility into unavailable package actions, additional requestor information across list/query/approve workflows, and expanded metadata passed to exits to support automation, auditing, and integration processes.Db2 Development ImprovementsWe’ve delivered several targeted updates that make Db2 change management smoother and more

Just for your information*)In UF 9, I had stored a list with GOLD- as a constant. That worked well.In UF 10, constants are created textually in DEFPARAM, and the semicolon acts as a comment character.The nice list is ruined :-(*) Even if Uniface were to correct this, it would take forever for it to be delivered in a version.“Unfortunately”, I need it right now. As a workaround, I used a pipe symbol instead of GOLD-;.And, of course, I have to convert | back to GOLD-; wherever it is used.

I’m trying to configure the db.mvbasic.json file for includeMapping.I’m wondering if the issue I’m having is because in our programs, our $INSERT lines do not contain the file name.  Somehow, don’t ask me how, regardless of which file the $INSERT is in, the compiler finds it and adds it.For example, here’s a typical line:     $insert ENVIRON_CONSTANTS.INS.IBASENVIRON_CONSTANTS.INS.IBAS is not contained within an actual Universe account, nor is it in the same file as the source code.  This is confusing me when configuring the includeMapping section.Here’s what I tried:I added my development account to accounts section:    "accounts": [        {            "name": "devbtr",            "path": "/home/dev/devbtr"        }    ],ENVIRON_CONSTANTS.INS.IBAS is located under /home/dev/cadsrc/INSERTIn the account devbtr, the VOC entry that points to it is CADSRC.INSERT     CADSRC.INSERT0001 F0002 /home/dev/cadsrc/INSERT0003 /uvdata/uv/D_UFD/home/dev/cadsrc is NOT a Universe account. I tried:   

DownloadsProduct availability Matrix (PAM)Downloads of this patch are available from the Rocket Customer Support Portal. New Feature(s)Logging on Unix is now faster and more efficient. Uniface optimizes $putmess_logfile and $transcript_logfile for multi‑core CPUs using multi‑threaded logging, resulting in improved application responsiveness and overall performance.   The label text defined on the modeled field is now exposed as a property for widgets created with the UX Interface, and the standard Uniface UX widgets have been updated to display it. This enhancement improves inheritance behavior and streamlines modernization efforts. The widget’s _label‑text_ property can be used to override the field's label at runtime. Classic DSP widgets will continue to work as before, using a dedicated Label object. Please refer to the user documentation for additional details.CurrenciesAdded Support for Oracle 23ai (U7.7) on 64-bit Windows Server 2025.

I’m told that universe support says: . ... they receive very few support cases related to NLS…There are a couple of possible reasons for this, one being that no-one actually uses it. We have code that supports some high-range ascii characters that would otherwise cause problems with the item storage but are looking to test a proper unicode solution. I really don’t want to be the first down this path ;-) Mike

Does anyone have an accurate handle on how and when AUTOLOGOUT triggers?   The manual says….Use AUTOLOGOUT to enable or disable automatic logout. With AUTOLOGOUT enabled, UniVerse logs you out automatically if you have not pressed a key within a specified time.But I don’t believe that is accurate.   We’ve had some confusing results in testing.   A process on a loop with periodic TIMEDATE() displays (output only) does not trigger.  Similarly a loop process with no screen output but just a SLEEP X also does not trigger even in between loops before it goes back to sleep.So, we are trying to understand the actual nuts and bolts of how/when it triggers.   Thanks for any help.Steve DouglasVP Systems AdminMBMS, LLC

If anyone finds themselves in the same situation as I was after applying patch 26 to 10.4.03 and the Ctrl+F no longer brings up the find box in the proc editor, then just add this to your usys.ini and it will get you going again. My guess is that when they added the feature “Go to line number in code” they broke whatever the defaults that were baked into the proc editor because the search and the search and replace worked just fine without having to be defined in the usys.ini file(s). I opened a support ticket on this and eventually solved it myself. Hopefully, they address it or at a minimum, add it automatically to the usys.ini in future patches as they are applied.[userkeys]code.search=Ctrl+Fcode.searchandreplace=Ctrl+Hcode.gotoline=Ctrl+G

This pertains to the use of external filehandlers, where my COBOL class is compiled with CALLFH"abcdef", which in tern needs to call fh/aaaaaa.dll which exists in the sub folder fh.The COBOL has :CALL “fh/aaaaaa” using ………During animation I debug the class, debug abcdef, but when it gets to the CALL “fh/aaaaaa” it doesn’t raise an exception but doesn’t execute the code either.I can see from procmon that the assembly was indeed loaded as well as the .pdb but just ignored almost as though it was looking for the programid but that didn’t match.Is this not supported?Is there an COBOL .NET  replacement for COBPATH ?Neil 

​​​​​​I am currently launching a Python script from RM/COBOL with CALL SYSTEM. Seems there is some overhead associated with SYSTEM, I am looking for a more efficient method.  Are there any alternatives recommended? 

We have a web page built in against an HTML widget in our CS. Clicking a button on the web page uses a JS trigger to start loading a CS form and populating it with data. Whilst this is happening (can be up to 10 seconds before the entire form is loaded/shown) the web page is active and there is no indication at the client end that anything is happening. I’d like to turn on the ‘Processsing’ function, as per the refresh/loading of the page in the first place, from JS calls to the HTML widget before the activate of the triggered form, and off again after the form is loaded. (So the user doesn’t think nothing is happening and click the button several times (causing multiple instances of the form to fire up). Is this possible? Putting the DSP into ‘locked/processing’ state without the round trip to the server refresh… Iain

I'm submitting a job submisison request to JESControl using the REST client to the following url  POST "">ip.address.of.server:10086/.../jescontrol" But my request is getting rate limited and I'm getting the following response: { "ErrorTitle" : "Error 429 - Too Many Requests", "ErrorMessage" : "Session limit has been reached. Please try again later."} How do I handle this response. Is the amount to wait before sending the next request is available in any HTTP header ? Usually when rate-limiting is implemented, the amount of seconds to wait will be available in the 'retry-after' header, however for the above request, I did not see any such header in the response received. So wanted to know if Enterprise Server JESControl is setting this value in any other header or if there is any other way to handle this.  Thanks for your response. Really appreciate.#EnterpriseServer

All of my requests being sent via RESTful API that previously worked have recently been receiving a 429 HTTP status code indicating too many requests have been sent to the API.I'm getting the following response:{"ErrorTitle" : "Error 429 - Too Many Requests","ErrorMessage" : "Session limit has been reached. Please try again later."}The error is a general HTTP message, not specific to the product.How do I mitigate this obstacle? What is the admin configuration setting in the Enterprise Server to extend the session limit and relax the rate-limiting?The Retry-After header is an HTTP response header that informs the client how long it should wait before making another request is commonly used in scenarios where the server cannot process the request immediately, such as rate-limiting or scheduled downtime. I do not want any of the rate-limiting or scheduled downtime that I’m currently experiencing.

I have a JCL script that executes a COBOL program through a procedure (PROC0001). Specifically, the goal is to pass parameters to the procedure so that any COBOL program can be executed, and this is source code that is running in a production environment. However, when I try to run the same code on my PC with Enterprise Developer for Eclipse 11 on a Windows platform, the following errors occur:But when running the JCL script, a series of errors occur.Specifically, here is the relevant portion of the JCL script://JCL0001 JOB (0),'JCL0001',CLASS=?,MSGCLASS=?,USER=????????//STEP1 EXEC PROC=PROC0001,// PROGRAM=PGM00001,// DSN=DBP,// PLAN=PPGM0001,// LOAD1=PGMS.LIB.LOADHere I attach the part of PROC PROC0001 that I find relevant://PROC0001 PROC PROGRAM=XXXL9999,// DSN=XXX,// PLAN=XXXP9999,// LOAD1=STEC.LIBDUMMY.LOAD,// LOAD2=STEC.LIBDUMMY.LOAD,// LOAD3=STEC.LIBDUMMY.LOAD,// OUTC=*,// RGN=0K,// PARMS=&PARMSAnd the result of the execution produces the following errors:1 //JCL0001 JOB (0),