>You know, it really ticks me off when you talk about avoiding portions of the language that are not UI-centric in a COM object, just to make excuses for Microsoft's bugs.
>

First, it's pretty well documented in any of a number of places that AMEMBERS() works with VFP-native objects. I'd always accepted this, and was surprised that Nancy found the behavior to be different than what I had thought. Maybe a side effect or an enhancement that hadn't been publicized. Or maybe I was just flat wrong all along!

I understand that you wouldn't be so foolish as to be wrong and less than omniscient, but, that price one must pay for righteous anger.

When we started having different results among three people who have some slight clue about how VFP works, we started investigating what was different; maybe something Nancy had loaded, or a configuration option difference, could give us some added functionality. And so the discussion began, comparing things, and in the end, with some input from a couple of people who know considerably more than we do about the innards of VFP, we all agreed, that AMEMBERS() couldn't reliably be used to probe ActiveX objects. No big deal; it wasn't supposed to, and the possibility of reliably configuring a system to make AMEMBERS() behave in a way we wanted, but that it never was supposed to do, fell through.

And you're peeved that we couldn't make VFP misbehave to fit our wishes? Sheesh.

And you're annoyed that George accepted that it didn't have an unannounced capability? Give me a break.

I suppose that recognizing that there is no universal cure for what results in a C5 error makes me an instrument of the evil empire...