Re: Troubleshooting a stalled procedure

[Scott Klement <rpg400-l@xxxxxxxxxxxxxxxx>]

Problems like this are very frustrating and hard to track down.  Try 
inserting some harmless that runs at the start of the WS3GETRGDV 
procedure, and see if that code runs.  (Have it send you a message or 
something.)

If that still fails, then I'd try commenting out code one line at a time 
until you figure out which line is the culprit.

Most likely you are corrupting something in memory or on the call stack 
that is causing things to go absolutely haywire.  Unfortunately, when 
stuff like this happens, I've found the debugger to be very unreliable, 
often skipping the breakpoints entirely.



On 4/4/2013 11:19 AM, Koester, Michael wrote:
> I'm searching for ideas on where to look to find what I need to correct.
>
> Environment:  IBM i at 7.1	WDSCi 7.0.0.8 (7.0.0.20090225_1508) 	PTF's are usually pretty current
>
> Brief description:  I've created a couple modules (WS3PRVSN and WS3PRVSN05) that comprise the "plumbing" to retrieve data from a web service.  A web service driver module (WS3GetRgDv) loads the input data structure, executes the call to the web service, and parses the results from the data structure returned by the web service.  These three modules are in a service program.
> A test program (TESTSTUFF) was created to pass the inputs to the web service driver.
>
> Issue: The test program calls the web service driver procedure, the SEP debugger shows that the called procedure loads, but hangs without advancing to the first executable statement.  WRKACTJOB show status RUN; call stack confirms that the intended procedure is at stmt 2100000001 (whatever that is), but nothing else happens.  I expected the first executable statement (one that tests for the presence of an optional parm) to execute, but the debugger doesn't get there.
>
> The test program runs in activation group, *NEW, and the service program was created to run in *CALLER.  All this is in my development environment, so I'm not in contention with anyone else.
>
> I've confirmed that all the modules, the service program, and the calling program are freshly created.  WRKOBJ shows everything appears to be in place.
> I call TESTSTUFF from a command line of a newly-started 5250 session, so there shouldn't be any zombies from prior attempts causing trouble.
> I've confirmed through stub testing that the web service plumbing programs are sound and return results with the same inputs my test program uses.  That shouldn't be an issue at this point, because I can't seem to get that far.
>
> Ideas?
>
> Thanks.
> Michael Koester