On Tue, Mar 22, 2022 at 3:01 PM Andres Freund <andres@anarazel.de> wrote:
> We clearly already know how to compute whether a lock is "included" in
> something we already hold - after all ProcSleep() successfully does so.
>
> Isn't it a pretty trivial test? Seems like it'd boil down to something like
I don't mind you fixing the behavior. I just couldn't pass up an
opportunity to complain about the structure of lock.c.
--
Robert Haas
EDB: http://www.enterprisedb.com