> 8) Master Scroll events fired
> 9) Detail Scroll events should NOT fire (result dataset
> STILL contains 0 records, hence it has not scrolled)

But it has scrolled, indeed it sort of scrolled twice :-)

1. The detail is closed as new master parameters are applied
so the dataset goes from having selected Eof to being
closed.

2. The dataset is opened as the new parameters are executed
so the dataset goes from being closed to having selected
Eof (which is what generates the scroll event in this
case I think).

AFAIK that is what is actually happening in the code. I suspect
that the situation is ultimately the result of code reuse (the same
class used for master and detail datasets) rather than necesarily an
intentional effect. However my personal opinion is that this is the
way it should operate; the detail is going from one empty selection
to another empty selection, they are different empty sets.

I can see you point about the detail still having selected an empty
record - physically the same record/state. But I see this as
coincidence rather than something that events should consider.

And again I would warn against changes in this regard, since it
is possible that other develop code exists that depends on this
aspect of IBO operation.


--
Geoff Worboys
Telesis Computing