scottklement.com

I have always had issues with OVR and SQL. Given that you know the name of the library you need why not use SET SCHEMA - if SQL ignores that you really have an issue! Another option would be to pass the required library name to the function and use it to qualify all table references.

I don't see how it could work when you specify the cell address range as Interval like(Jstring). It has to refer to an object of class CellRangeAddress. You don't show your proto for CellRangeAddress.

No idea why you are stuck on 7.2 but ...

I still think you need to look at the formation of the request and/or the headers. The response I got - using your exact request - was that the request was mail-formed.

Can't help a lot with this but I noticed that switching from the old ** SYSTOOLS.HTTPgetCLOBverbose to QSYS2.HTTP_GET allowed me to get the
response back, which clearly states that: "That’s an error. Your client has issued a malformed or illegal request. That’s all we know."

One of the posts here ...

Well - according to Scott's site, the latest version runs on 6.1, so you should have no issues. See https://www.scottklement.com/httpapi/

Where most people encounter issues with out-of-date releases is that useful RPG features like DATA-INTO and DATA-GEN may not be available.

Your basic problem is that you are treating return_message as an object but it is an array. So you need to work through it using YAJL_array_loop and _then_ YAJL_OBJECT_LOOP for its elements. It appears that return_message is a single element so heaven knows why the doc was designed like this but ...

I'll try to look at your code later, but is there a reason why you are using the "raw" YAJL APIs? I just ran your JSON through YAJLGEN, and the resulting DATA-INTO works perfectly on this document.

It always saddens me when programmers are forced to find "solutions" for non-existent problems. The answer is so easy - update to a supported release.

But if the powers that be deem this impossible the only thiong I can think of is to load a batch file on each PC and invoke that with parms for the ...

Scott is correct - but these days there is another option which is to use the ability to define the data type and length of parms on a call. Like so:

CALL PGM(XYZ) PARM(('ABC' (*CHAR 100)))

In this case, the system will pass a 100-character field blank-padded as expected.

That limit disappeared back in release 7.2. Are you running a really old release? The current limit is 1,023 characters. Of course you will also need to be running a 5250 emulator package that is aware of this update.