Re: EXEC_BACKEND vs bgworkers without BGWORKER_SHMEM_ACCESS

On Mon, Aug 09, 2021 at 11:07:14AM -0400, Robert Haas wrote:
> On Thu, Aug 5, 2021 at 8:02 PM Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> I think doing nothing is fine.  Given the lack of complaints, we're
>> more likely to break something than fix anything useful.
> 
> +1.

FWIW, the only interesting case I have in my plugin box for a
background worker that does not attach to shared memory is a template
of worker able to catch signals, to be used as a base for simple
actions.  So that's not really interesting.  Making the SHMEM flag be
something mandatory on HEAD while doing nothing in the back-branches
sounds good to me, so +1.
--
Michael