On Tue, May 28, 2019 at 6:57 AM Amit Langote <langote_amit...@lab.ntt.co.jp> wrote: > > On 2019/05/27 21:56, Tom Lane wrote: > > Amit Langote <langote_amit...@lab.ntt.co.jp> writes: > >> On 2019/05/24 23:28, Tom Lane wrote: > >>> So my thought, if we want to do something about this, is not "find > >>> some things we can pfree at the end of planning" but "find a way > >>> to use a separate context for each statement in the query string". > > > >> Maybe like the attached? I'm not sure if we need to likewise be concerned > >> about exec_sql_string() being handed multi-query strings.
the whole extension sql script is passed to execute_sql_string(), so I think that it's a good thing to have similar workaround there. About the patch: - * Switch to appropriate context for constructing querytrees (again, - * these must outlive the execution context). + * Switch to appropriate context for constructing querytrees. + * Memory allocated during this construction is released before + * the generated plan is executed. The comment should mention query and plan trees, everything else seems ok to me.
