Masahiko Sawada <sawada.mshk@gmail.com> writes: > I don't think we need additional PG_TRY() for that since exec_stmts() > is already called in PG_TRY() if there is an exception block. I meant > to call stmt_end() in PG_CATCH() in exec_stmt_block() (i.e. only when > an error is caught by the exception block). Currently, if an error is > caught, we call stmt_begin() and stmt_end() for statements executed > inside the exception block but call only stmt_begin() for the > statement that raised an error.
I fail to see anything wrong with that. We never completed execution of the statement that raised an error, but calling stmt_end for it would imply that we did. I think changing this will break more things than it fixes, completely independently of whatever cost it would add.
Or in other words: the initial complaint describes a bug in pg_hint_plan, not one in plpgsql.
The OP suggests needing something akin to a "finally" callback for statement. While a fine feature request for plpgsql its absence doesn't constitute a bug.