Re: Autogenerate some wait events code and documentation

From: Noah Misch <noah(at)leadboat(dot)com>
To: Bertrand Drouvot <bertranddrouvot(dot)pg(at)gmail(dot)com>
Cc: Michael Paquier <michael(at)paquier(dot)xyz>, PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: Autogenerate some wait events code and documentation
Date: 2024-03-18 15:49:34
Message-ID: 20240318154934.10.nmisch@google.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Mon, Mar 18, 2024 at 09:04:44AM +0000, Bertrand Drouvot wrote:
> --- a/src/backend/utils/activity/wait_event_names.txt
> +++ b/src/backend/utils/activity/wait_event_names.txt
> @@ -24,7 +24,12 @@
> # SGML tables of wait events for inclusion in the documentation.
> #
> # When adding a new wait event, make sure it is placed in the appropriate
> -# ClassName section.
> +# ClassName section. If the wait event is backpatched to a version < 17 then
> +# put it under a "Backpatch" delimiter at the end of the related ClassName
> +# section.

Back-patch from v17 to pre-v17 won't use this, because v16 has hand-maintained
enums. It's back-patch v18->v17 or v22->v17 where this will come up.

> +# Ensure that the wait events under the "Backpatch" delimiter are placed in the
> +# same order as in the pre 17 wait_event_types.h (see VERSION_FILE_SYNC as an
> +# example).

I expect the normal practice will be to put the entry in its natural position
in git master, then put it in the backpatch section for any other branch. In
other words, the backpatch regions are always empty in git master, and the
non-backpatch regions change in master only.

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Tomas Vondra 2024-03-18 15:55:12 Re: BitmapHeapScan streaming read user and prelim refactoring
Previous Message Ashutosh Bapat 2024-03-18 15:47:31 Re: Memory consumed by child SpecialJoinInfo in partitionwise join planning