Marko Ristola wrote:
> I have been thinking about savepoints.
> What do you think about the following:
The interface and behavior of the ODBC driver is bound by various public
standards such as the SQL standard and the Microsoft ODBC
pseudostandard. So any new functionality should be based on either of
these sources. Does that apply to your proposal?
> They would keep the knoledge about the following things:
> - Current transaction state (is in autocommit mode, in transaction,
> and the savepoint stack structure.
This information is available on the protocol level and does not to be
tracked.
> This would make the savepoint rollback work with cursors:
I'm not exactly up to speed on how savepoints interact with cursors
nowadays. Perhaps you could first explain what problem you are trying
to solve?
--
Peter Eisentraut
http://developer.postgresql.org/~petere/