'Morning all,

Well I just 'solved' a real strange one and thought it would be useful to share it...

Starting Friday I began testing some significant changes to a WebConnect process (2 actually, rather large).
Early testing looked very positive except for a coupla fields that had slightly erroneous content. Examination showed that an IF statement evaluating a VAL(...) was either not executing or giving an incorrect result and two STRTRAN statements weren't executing. There may have been other stuff but these are what stood out.No error messages or other signs of trouble - just wrong content in the output table.

Anyway, try as I might, I couldn't get this stuff to deliver the correct output. Executing in the command window always gave correct results. Variations on the IF yielded nothing. Finally it was time to go home so I left it.

On my way in this morning I struck on a silly notion - the commands that seemed to non-execute were both functions late in the alphabet. I wondered if I might have exceeded some internal limit.
Now, as I often do, I had kept a copy of the original processes right in the same source, renaming the FUNCTION statement to accomplish this. The combined routines took, with commentary, well over 1000 lines.
When I got to the office I deleted the extraneous copies of the 2 routines, reran my test and, you guessed it - it ran perfectly!!!!

If anyone has some other logic to support what happened I'd love to hear it. I simply feel lucky to have acted on such a silly sounding impulse and wouldn't have had it involved any more work than it did.

Cheers,

JimN