> 3. The problem I have is that the transaction started by IBO when
> reading stuff from the system tables when the first query is
> opened, does not get closed as is the usual case. The result was
> that the 'oldest transaction' got stuck, leading to problems with
> garbage collection. The workaround (which I don't like) is to use
> the connection's 'Transaction[]' property to commit all
> transactions at a point in the software where I know there
> shouldn't be any open transactions.

This problem should be fixed (cant remember when and I cant see
reference to it in the release notes).

The schemacache is now setup to check that the internal transaction
is automatically closed if the associated session has a timer value
of 0 (which should mean automatically ended on all background
threads.)

--
Geoff Worboys
Telesis Computing