| From: | Michael Paquier <michael(at)paquier(dot)xyz> |
|---|---|
| To: | Kyotaro Horiguchi <horikyota(dot)ntt(at)gmail(dot)com> |
| Cc: | katouknl(at)oss(dot)nttdata(dot)com, pgsql-hackers(at)lists(dot)postgresql(dot)org |
| Subject: | Re: CREATE tab completion |
| Date: | 2021-11-19 04:21:48 |
| Message-ID: | YZcmXIaU8ZoBZwuz@paquier.xyz |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Thu, Nov 18, 2021 at 05:20:58PM +0900, Kyotaro Horiguchi wrote:
> So we could use COMPLETE_WITH_CS instead so that CREATE SEQUENCE
> behavess the same way with the existing behavior of TYPE.
Makes sense.
Another issue I have noticed with the patch is that it forgets to
apply quotes for the FROM/TO clauses of CREATE CONVERSION, making the
queries fail. \encoding does not care about that but we do for this
set of DDLs. So I have discarded this part.
One extra issue was in CREATE TRANSFORM, where we should handle "OR
REPLACE" like the others. And a last issue I had was with CREATE
LANGUAGE where we would miss TRUSTED. This last one cannot be
completed, but we'd better allow its completion if we can. That makes
things a bit tense if you add on top of that the handling of "OR
REPLACE".
I have fixed (or discarded) that, and the parts for sequences, domains
and transforms remained. That looked like good enough on its own, so
applied those parts of the patch.
--
Michael
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Bharath Rupireddy | 2021-11-19 04:24:39 | Re: Should we improve "PID XXXX is not a PostgreSQL server process" warning for pg_terminate_backend(<<postmaster_pid>>)? |
| Previous Message | Amit Kapila | 2021-11-19 04:18:10 | Re: row filtering for logical replication |