On 5 Mar 2019, at 3:22am, Rowan Worth <[email protected]> wrote: > Wait what? If I've understood correctly you're describing a situation where > statements outlive their transaction context?
The documentation on transactions covers this possibility. This includes continuing to do _step() for one statement while another executes END or ROLLBACK. Your guess about how it is treated is correct. Nevertheless I consider this behaviour extremely high-risk and would reject any production program which depended on it as the product of a poor programmer. Simon. _______________________________________________ sqlite-users mailing list [email protected] http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
