On Thu, Oct 24, 2024 at 5:44 PM Peter Smith <smithpb2...@gmail.com> wrote: > > On Thu, Oct 24, 2024 at 5:41 PM Michael Paquier <mich...@paquier.xyz> wrote: > > > > On Thu, Oct 24, 2024 at 05:27:25PM +1100, Peter Smith wrote: > > > Yes, well spotted -- I was aware of that. Originally I had coded a >= > > > PG15 check for that pub_names assignment. e.g. > > > > > > if (walrcv_server_version(LogRepWorkerWalRcvConn) >= 150000) > > > GetPublicationsStr(MySubscription->publications, pub_names, true); > > > > I was wondering about putting the call of GetPublicationsStr() in the > > first block with the makeStringInfo(), have an Assert checking that > > pub_names->data is never NULL in the second block, and destroy the > > StringInfo only if it has been allocated. > > > > > But, somehow it seemed messy to do that just to cater for something I > > > thought was not particularly likely. OTOH, I am happy to put that > > > check back in if you think it is warranted. > > > > I think I would add it. The publication list is not mandatory, but > > your patch makes it look so. > > -- > > No problem. I will post an updated patch tomorrow. >
I've attached the patch v4. ====== Kind Regards, Peter Smith. Fujitsu Australia
v4-0001-Refactor-to-use-common-function-GetPublicationsSt.patch
Description: Binary data
