Re: Odd query execution behavior with extended protocol

Robert Haas <robertmhaas@gmail.com> writes:
> From looking at the code, it appears to me that if the Execute is run
> to completion, then its results will be seen by future statements, but
> if the portal is closed earlier, then not.  See the end of
> exec_execute_message.  The handler for the Close message (inside
> PostgresMain) doesn't seem to do anything that would result in a
> CommandCounterIncrement() or CommitTransactionCommand().

> This does seem a little strange.

I dunno, if you close a portal before you've gotten CommandComplete,
should you expect that all its actions were taken?  Arguably, that
should be more like a ROLLBACK.

Note there'd only be a difference in case of an operation with RETURNING,
else there's no way (at this level anyway) to pause a data-modifying
command mid-execution.  This logic all predates RETURNING, I think,
so maybe it does need to be revisited.  But it's not entirely clear
what should happen.
        regards, tom lane