Tom Marchant wrote: | I am not surprised by the behavior that | you describe.
and I should be disagreeably surprised by any other behavior. This is an old problem (if that is what it is). If you wish to preserve a current return code you must save it before and restore it after any subroutine call. Anciently, when only powers of 2 were used as return codes in a disciplined way, it was possible to set them additively in a fashion that permitted multiple errors/anomalies to be diagnosed in a single return-code value, but this scheme was always a fragile one, and it is no longer usable outside a small, closed subsystem. John Gilmore, Ashland, MA 01721 - USA ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
