I'm pulling my hair out trying to figure this one out. Today I got an error message indicating non-numeric data was used where numeric data is required in program RTP8. Well, RTP8 is WRITEV, so I checked the calling program for WRITEV calls. There is only one. The last parameter of a WRITEV call is a number indicating the field position to write to. In this case, it is a constant equated to 1238 (that is not a typo). The rest of the call consists of a file handle, a string, and an id, none of which require numeric entries. So what is causing the error? Interestingly, the problem does not recur when the program is restarted.

Stranger still is another program that occasionally breaks, always at the same place. It is the same error message, and the break always occurs at the same place: a WRITE statement. I couldn't believe the message was accurate given the line of code indicated, so I recompiled the source program. I even added a blank line for good measure to see if the problem would still occur at the same statement. It did.

In each time, when I query the variables in question from the debugger, they look correct for the context. This is really aggravating because it occurs during what is supposed to be unattended night-time batch processing. A major consequence of this is that I am losing a lot of sleep occasionally babysitting the batch jobs in case they abort.

Any solutions or things to try greatly appreciated.

Victor

Any ideas?

most likely a bug in an underlying MFS - never reports the MFS line number just the program line number.

amcauley@sprezzatura.com

Sprezzatura Ltd

World Leaders in all things RevSoft

I was wondering about that. Thanks for confirming my suspicion. The missing piece of information is why it didn't report the MFS name.

View this thread on the forum...