Re: rtsora broken??

nobody.noclan_at_bigfoot.com (Eric) wrote
>
> Thanks for the thoughts, but I can't see it being the runtime
> environment. There are many other programs procoboled and compiled in
> the same way which are running successfully. The problem seems
> specific to this program, and only when this program is called as a
> subroutine of a subroutine.

Hmm.. in that case you could be running out of stack.. but that's something I last ran into with 16bit developmemnt using a lot of recursion... Is the working storage particularly large?

I would try and isolate the problem by trying to test this one step at a time. Make this program an empty stub that does nothing. Call it as a subroutine from a subroutine and see if that works. If it does, slowly add more code to it testing all the way (eventually pre-compiling it) until the error is encountered.

Another thing to check is the pre-cobol output. I once had a very weird error in a Pro*C program that did batch processing. I finally found the error when looking at the C source output from the pre-compiler. The SQL statement was too large (had a lot of unions in it). Pro*C catered for large SQL statements by only doing two iterations of a certain set of OCI calls. But this failed to cover the complete SQL statement. As luck would have it, it cut off the SQL statement at a WHERE clause.. which meant that the SQL statement that was pre-compiled was valid during runtime, even when a huge chunk of it was missing.

So it's worthwhile going thru the pre-compiler source output and having a close look at that IMO.

--
Billy