Re: Hooks for session start and end, take two

From: Michael Paquier <michael(at)paquier(dot)xyz>
To: Fabrízio de Royes Mello <fabriziomello(at)gmail(dot)com>
Cc: legrand legrand <legrand_legrand(at)hotmail(dot)com>, Pgsql Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Hooks for session start and end, take two
Date: 2019-09-30 01:29:41
Message-ID: 20190930012941.GD2888@paquier.xyz
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Sat, Sep 28, 2019 at 10:43:36AM -0300, Fabrízio de Royes Mello wrote:
> I poked a little with the patch and everything is ok. Your check for normal
> backend on test_session_hooks is much simpler than I did before:
>
> + /* just consider normal backends */
> + if (MyBackendId == InvalidBackendId)
> + return;
>
> But one thing came to my mind, why not in this first version we hook just
> normal backends?

This code path can only be taken by normal backends, so that would
apply, still I don't actually see why we should limit us here on the
backend side. If for a reason or another those two code paths begin
to be taken by a backend with InvalidBackendId, then users of the
session start/end hook will need to think how to handle it if they
didn't from the start, which sounds like a good thing to me.
--
Michael

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Fujii Masao 2019-09-30 01:48:53 Re: Standby accepts recovery_target_timeline setting?
Previous Message Chapman Flack 2019-09-30 01:24:08 documentation inconsistent re: alignment