Re: WIP: a way forward on bootstrap data

From: Andres Freund <andres(at)anarazel(dot)de>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: John Naylor <jcnaylor(at)gmail(dot)com>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: WIP: a way forward on bootstrap data
Date: 2018-04-04 22:34:43
Message-ID: 20180404223443.dufuj4qo2lz2ychf@alap3.anarazel.de
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi,

On 2018-04-04 18:29:31 -0400, Tom Lane wrote:
> I'm starting to look through v13 seriously, and one thing struck
> me that could use some general discussion: what is our policy
> going to be for choosing the default values for catalog columns?
>
> [...]
>
> In short, I'm tempted to say that when there's a clear majority of
> entries that would use a particular default, that's the default we
> should use, whether or not it's "surprising" or "unsafe" according
> to the semantics. It's clearly not "surprising" for a C function
> to be marked proparallel 's'; the other cases are more so.
>
> [...]
>
> I'm not seeing any other BKI_DEFAULT choices that I'm inclined to
> question, so maybe it's a mistake to try to derive any general
> policy choices from such a small number of cases. But anyway
> I'm inclined to change these cases.
>
> Comments anyone?

I think choosing SQL defaults is defensible, but so is choosing the most
common value as default to make uncommon stand out more, and so is
choosing the safest values. In short, I don't think it matters terribly
much, we just should try to be reasonably consistent about.

Greetings,

Andres Freund

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message legrand legrand 2018-04-04 22:43:52 Re: [PROPOSAL] timestamp informations to pg_stat_statements
Previous Message Tom Lane 2018-04-04 22:29:31 Re: WIP: a way forward on bootstrap data