>>>>
>>>>In VFP7, does it fire the error on a table that contains a field named IRegion? IOW, did you run a COUNT FOR x=y statement against a table where the left side (x) was valid, but the right side (y) was invalid?
>>>
>>>Sorry, I was wrong. It works the same way as in VFP 6.
>>
>>Part of the craziness is caused when the table is empty...the error does not fire because VFP does not need to inspect the FOR clause.
>
>Yes, well I have to say that if the error does not fire when the table is empty, then it is ok.
>
>Now, it's just a question of making sure the code does not proceed on an error and have the RETURN TO to work.

Are you getting any closer to getting the RETURN TO to work? I played with it some more, it just doesn't make much sense.

It is almost like the call stack is blown by the error in the COUNT command. However, if you inspect the call stack in the debugger in the error method (or if you dump it out with calls to PROGRAM()), it appears normal.