D
dbeavon
Guest
I'm trying to wrap my head around that. So my legacy handling of the STOP condition is "DO ON STOP UNDO, RETURN ERROR v_UnexpectedError". You are saying that this type of STOP handling always sends the related error messages to the logs. But if I use "-catchStop 1" and put a CATCH statement in there (for Progress.Lang.LockConflict) then you would expect that messages would no longer be sent to the logs. Am I getting that right? Conceptually the two don't seem that different to me. They are both handling the STOP condition that is arising from a lock conflict. I suppose that in the case of the "ON STOP" phrase, the ABL programmer doesn't have any of the related details (and there is no way to get at them). But in the case of the CATCH statement the ABL programmer has all of the necessary details and they can determine for themselves whether or not they want to write them to a log file. I was under the impression that the "traditional error handling model" would also suppress messages from the logs if you did certain tricks like using "NO-ERROR" after various types of statements (and proceeded to inspected the "error-status system handle" after the fact). Maybe the NO-ERROR trick is a slightly different topic. It appears that the NO-ERROR trick might have more of an impact on the agent logs than using "ON STOP" or "ON ERROR" phrases on a DO block.
Continue reading...
Continue reading...