Alvaro Herrera <alvherre(at)dcc(dot)uchile(dot)cl> writes:
> What's needed for this is to isolate the transaction-initiating code
> from the main query-processing loop. So for CALL statements it wouldn't
> be invoked, and the procedure would be able to use its own explicit
> transaction blocks and savepoints.
Is that really needed? What if the procedure starts in a transaction normally
but is just allowed to commit it and start another transaction? I mean it's
not like it would be allowed to do any actual work without starting a
transaction anyways.
--
greg