Consider the following:
:::cobolfree
DISPLAY "blah" AT COL 100000000000000
ACCEPT OMITTED
The DISPLAY statement will cause EC-SCREEN-STARTING-COLUMN to exist and will cause FUNCTION EXCEPTION-STATUS to return that exception condition. However, the ACCEPT resets the last exception, causing FUNCTION EXCEPTION-STATUS to return a blank string afterwards. I believe similar behaviour occurs after calling some intrinsic functions.
This should not happen as EXCEPTION-STATUS returns the name of the last exception raised, not one that happened to be raised by the previous statement. The last exception status may only be blanked by an explicit SET LAST EXCEPTION TO OFF statement.
Very good point! Any chance for you to look at this?
Yes, but not immediately. I'll do some of the stuff from [feature-requests:#103] first. You're welcome to work on this, if you want.
Related
Wish List:
#103Although it looks like a quick victory (don't set the exception to zero despite on runtime init and
SET LAST EXCEPTION TO OFF) I'll stick to all the other points already on my agenda.But who knows what happens next...
Whoever starts to investigate for necesary changes leaves a note here and we're fine.
I've added
SET LAST EXCEPTION TO OFFin [r831].Fixed by [r2303].