| From: | Peter Eisentraut <peter(at)eisentraut(dot)org> |
|---|---|
| To: | pgsql-hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | doc: should pg_createsubscriber be grouped as a client application? |
| Date: | 2026-06-22 06:39:57 |
| Message-ID: | 1adc75e6-8b8f-4e53-ba32-a8cd0ca38b6c@eisentraut.org |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
pg_createsubscriber is currently grouped as a Server Application:
https://www.postgresql.org/docs/devel/reference-server.html
Contrast the list of Client Applications:
https://www.postgresql.org/docs/devel/reference-client.html
Note that the latter includes pg_basebackup, pg_receivewal, and
pg_recvlogical, which feel generally similar to pg_createsubscriber.
My mental model for Server Applications is "does not make a remote libpq
connection", which does not match what pg_createsubscriber does.
(An indirect effect of this grouping is how packagers split their
-client and -server packages. There are currently a number of
discrepancies among popular packagings, including this one. (I know
that moving files between packages is sometimes difficult, and I don't
necessarily expect that packagers would immediately move this, but at
least the directional guidance could be made more correct.))
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Chao Li | 2026-06-22 06:54:35 | Re: [PATCH] Add pg_get_table_ddl() to reconstruct CREATE TABLE statements |
| Previous Message | Richard Guo | 2026-06-22 06:37:04 | Re: assertion failure with unique index + partitioning + join |