On Wed, Mar 29, 2023 at 7:44 AM Peter Smith <smithpb2...@gmail.com> wrote: > > A minor review comment for v25-0001. > > ====== > src/backend/commands/subscriptioncmds.c > > 1. > @@ -1936,21 +1936,56 @@ fetch_table_list(WalReceiverConn *wrconn, List > *publications) > WalRcvExecResult *res; > StringInfoData cmd; > TupleTableSlot *slot; > - Oid tableRow[3] = {TEXTOID, TEXTOID, NAMEARRAYOID}; > + Oid tableRow[3] = {TEXTOID, TEXTOID, InvalidOid}; > > The patch could be slightly less invasive if you did not make this > change, but instead, only overwrite tableRow[2] for the >= PG16 case. > > Or vice versa, if you prefer. >
The current coding pattern looks neat to me. -- With Regards, Amit Kapila.