> In my view if it isn't a bug it's an unacceptable consequence
> of the way IB works.

Thats one to take up on the IB lists - presumably IB_Architect -
however you may want to hold off to see if Jason has anything else to
add.

> With regard to your suggestion for a full fix - how about
> the TIB_Dataset doing a refetch only if there is more than
> one blob field in the record?

Two things...

1. A second cursor containing a blob (on the same transaction) will
cause the same problem as two blobs in the one cursor.

2. ANY automation in this regard could significantly impact others
applications. TIB_Cursor is often used to step through data as in...

open cursor
while not eof
do something
post
next
end

If IBO started to automate the refresh then such processes would take
significantly longer because each post may result in a refresh.


I think the only viable solution is for this to be documented - so
that anyone using TIB_Cursor as you are trying to do, is reminded to
do a manual refresh after a post.

May not be perfect, but we can only do what we can :-)

Geoff Worboys
Telesis Computing