Re: Little confusing things about client_min_messages.

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Tomonari Katsumata <katsumata(dot)tomonari(at)po(dot)ntts(dot)co(dot)jp>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Little confusing things about client_min_messages.
Date: 2014-03-10 14:45:50
Message-ID: 32603.1394462750@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Tomonari Katsumata <katsumata(dot)tomonari(at)po(dot)ntts(dot)co(dot)jp> writes:
> Adding FATAL and PANIC to client_min_messages is done at below-commit.
> 8ac386226d76b29a9f54c26b157e04e9b8368606
> http://git.postgresql.org/gitweb/?p=postgresql.git;a=commit;h=8ac386226d76b29a9f54c26b157e04e9b8368606

> According to the commit log, it seems that the purpose
> is suppressing to be sent error message to client when "DROP TABLE".
> In those days(pre 8.1), we did not have "DROP IF EXISTS" syntax,
> so it was useful.

> If this was the reason, now(from 8.2) we have "DROP IF EXISTS" syntax,

Uh, that was one example of what it might be good for; I doubt that the
use-case has now vanished entirely. While I'm still dubious about the
reliability of suppressing error messages, if people have been using this
type of coding for nearly 10 years then it probably works well enough
... and more to the point, they won't thank us for arbitrarily removing
it.

I think we should leave established practice alone here. It might be
confusing at first glance, but that doesn't mean it's the wrong thing.

regards, tom lane

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Robert Haas 2014-03-10 14:48:20 Re: Retain dynamic shared memory segments for postmaster lifetime
Previous Message Tom Lane 2014-03-10 14:33:25 Re: calculating an aspect of shared buffer state from a background worker