Re: Getting rid of SQLValueFunction

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es

From: Michael Paquier <michael@paquier.xyz>

To: Tom Lane <tgl@sss.pgh.pa.us>

Cc: Corey Huinker <corey.huinker@gmail.com>, Postgres hackers <pgsql-hackers@lists.postgresql.org>

Date: 2022-10-25T05:20:12Z

Lists: pgsql-hackers

Attachments

v3-0001-Remove-from-SQLValueFunction-all-the-entries-usin.patch (text/x-diff) patch v3-0001
v3-0002-Replace-SQLValueFunction-by-direct-function-calls.patch (text/x-diff) patch v3-0002

On Fri, Oct 21, 2022 at 02:27:07PM +0900, Michael Paquier wrote:
> I have looked at that, and the attribute mapping remains compatible
> with past versions once the appropriate pg_proc entries are added.
> The updated patch set attached does that (with a user() function as
> well to keep the code a maximum simple), with more tests to cover the
> attribute case mentioned upthread.

Attached is a rebased patch set, as of the conflicts from 2e0d80c.
--
Michael

Commits

Fix precision handling for some COERCE_SQL_SYNTAX functions
- 7aa81c61ec8f 16.0 landed
Replace SQLValueFunction by COERCE_SQL_SYNTAX
- f193883fc9ce 16.0 landed
Switch SQLValueFunction on "name" to use COERCE_SQL_SYNTAX
- fb32748e32e2 16.0 landed
Improve coverage of ruleutils.c for SQLValueFunctions
- 2e0d80c5bb72 16.0 landed