.LIS files not being created properly (found the problem!)

[Seeley Scott <Seeley_Scott@PRC.COM>]

Hi all,

Last week I posted with the following problem:

> I have a problem that may be caused by a .LIS file not being created
> properly.  This SQR report was written to work under PeopleSoft 6.0
> and it runs FINE when run on the client (a Windows NT ascii file),
> either through Peoplesoft or via SQR Workbench.  This same SQR was
> then copied up to Unix (to run on the server), but it blows up there.
>
> KA-BOOM!
>
> Actually it's not that exciting.  There's no error message even, it
> just runs to the end but never exits.  Stuck.  The Unix "kill" command
> finally has to be used to stop the process.
>
> Wanna know exactly where the program hangs up?  Ahhh, good question.
> A SHOW command was put on the line just before END-REPORT.  It shows
> up in the sqr.log file - so we know the program is getting all the way
> to END-REPORT but not actually ending.  Weird, huh?  I'm guessing that
> the report runs, finishes, but when it tries to create the .LIS file
> file just before exiting, it pukes and locks up for some reason.
> Guessing.  This would make sense because when there's no print
> statements (and therefore no .LIS file to create) it runs fine.

> By commenting out all the print statements on the Unix file,
> everything runs peachy. We can see from a tasty SHOW statements in
> the SQR.LOG file that everything runs through the end. But add back
> one SINGLE print statement into the program, which we did - "print
> 'please run, you miserable expletive' (+1,,)" - and it goes ka-boom
> again.
>
> This SQR is very similar to at least 25 other SQR's running on the
> same system, written by the same person over the past few months, and
> using much of the same code. This is the ONLY one that has problems.
> And it only happens in Unix. PeopleSoft and SQR Workbench runs the
> SQR perfectly. There are no control characters in it that we can
> find. Since the other SQR's run fine so it can't be a configuration
> issue.


Okay, we found the problem, with the help of the many responses from
people on this list.  THANK YOU everyone!!  Ray Ontko turned out to have
the winning idea.  He wrote:

>Does your program have a BEGIN-HEADING or BEGIN-FOOTING section?
>If so, is it possible that one of these is generating an error?
  (snip)
>It sounds like the problem happens hen the first page of your report is
>generated, which made me think of the BEGIN-HEADING or
>BEGIN-FOOTING sections.  Try removing these and see if you still get
>the problem.


And that was it.  The heading called another proc (don't ask), which had
a line of code that was supposed to print a blank line:

   PRINT ' '   (+1,,)

Except there was no space between the single quotes.  It read:

  PRINT ''   (+1,,)

Aaaaaaagh!!!  It ran fine on the client.  Bombed out in Unix with no
error messages.  Half of me is laughing out loud and the rest wants to
trash my office.  Anyways, hope this helps anyone who runs into a
similar problem in the future.

Ray, we owe you a life-sized chocolate llama - where should we send it??
(Unfortunately, someone bit the head off, though...  WHY does that
always happen?!?!)

Okay, we're going live in 2 days, so luv ya'll go bye now.  :-)