Re: [PATCH] Add pg_get_subscription_ddl() function
solai v <solai.cdac@gmail.com>
From: solai v <solai.cdac@gmail.com>
To: Vaibhav Dalvi <vaibhav.dalvi@enterprisedb.com>,
Peter Smith <smithpb2250@gmail.com>
Cc: pgsql-hackers@lists.postgresql.org, alvherre@kurilemu.de
Date: 2026-05-15T04:45:28Z
Lists: pgsql-hackers
Hi all, I reviewed and tested the v8 patch on my current branch tree. The patch functionality works correctly as expected after resolving manual conflicts during patch application. After testing, I verified the function availability for both name and oid variants, generated CREATE SUBSCRIPTION DDL that appears syntactically valid and complete, handling of invalid and NULL inputs, permission enforcement for pg_read_all_data / pg_create_subscription, quoted identifier handling for subscription/publication names and disabled subscription reconstruction behavior. And tested logical replication setup with publisher/subscriber clusters and verified the generated DDL syntax appears correct and reusable. A few additional observations were also noted that the connect=false behavior matches the documented rationale and pg_dump behavior and the generated output currently emits all subscription options, including default-valued ones. Overall, the patch behaved correctly in my testing and looks good to me. Regards, Solai