There does not appear to be a combination of NUMPROC and OPT that will cause an abend.
We did some testing. We ended up passing the literal '86 ' in via the linkage section to short circuit the compiler's optimizations. We tried compiling with all the combinations of NUMPROC and OPT. We found that, so long as the sending (usage display) field was unsigned, we got no S0C7 on the move. We also tried adding a compute statement using the packed field and that also worked fine. We passed in spaces and low values and those worked too. Adding a sign to the sending field made the S0C7 happen on the move because the generated code was a pack followed by a zap. It appears that the compiler generates code to smudge the sign into correctness when the sending field is unsigned. This is documented in the language reference under 6.2.24.2 Elementary move rules. A bit of a surprise. Tomorrow I'll have to try this with OS/VS COBOL (the only other compiler we've got hanging around) and see what happens. Code is available on request. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
