| From: | Michael Paquier <michael(at)paquier(dot)xyz> |
|---|---|
| To: | Nathan Bossart <nathandbossart(at)gmail(dot)com> |
| Cc: | Bertrand Drouvot <bertranddrouvot(dot)pg(at)gmail(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
| Subject: | Re: Verify predefined LWLocks tranches have entries in wait_event_names.txt |
| Date: | 2025-07-23 03:23:37 |
| Message-ID: | aIBVuScop3uVGThI@paquier.xyz |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Tue, Jul 22, 2025 at 02:25:13PM -0500, Nathan Bossart wrote:
> I stared at this patch some more and came up with the attached. The
> biggest change is that I've moved the list of built-in LWLock tranches to
> the existing lwlocklist.h file. That simplifies the patch and centralizes
> these lists. This is arguably a bit too much preprocessor magic, though.
> Thoughts?
With the argument about checking the consistency of the data between
wait_event_names.txt, that looks like an improvement to me. And as
far as I can see, the check you are adding in generate-lwlocknames.pl
also makes sure that the ordering of the entries is correct.
In short, I side with the argument that this extra magic will save
cycles overall.
--
Michael
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Amit Kapila | 2025-07-23 03:35:06 | pgsql: Preserve conflict-relevant data during logical replication. |
| Previous Message | Michael Paquier | 2025-07-23 03:15:06 | Re: track generic and custom plans in pg_stat_statements |