From: Tom Lane <tgl@sss.pgh.pa.us>
To: Stephen Frost <sfrost@snowman.net>
cc: Peter Eisentraut <peter.eisentraut@2ndquadrant.com>,
        Alvaro Herrera <alvherre@2ndquadrant.com>,
        "imai.yoshikazu@fujitsu.com" <imai.yoshikazu@fujitsu.com>,
        Euler Taveira <euler@timbira.com.br>,
        Andrew Gierth <andrew@tao11.riddles.org.uk>,
        Justin Pryzby <pryzby@telsasoft.com>,
        PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Subject: Re: v12 and pg_restore -f-
In-reply-to: <20191105141132.GK6962@tamriel.snowman.net>
References: <20191104145336.GA23924@alvherre.pgsql>
 <5e124b1d-bd8f-0659-1088-4952464223bc@2ndquadrant.com>
 <20191105141132.GK6962@tamriel.snowman.net>
Comments: In-reply-to Stephen Frost <sfrost@snowman.net>
	message dated "Tue, 05 Nov 2019 09:11:32 -0500"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <7896.1572964109.1@sss.pgh.pa.us>
Date: Tue, 05 Nov 2019 09:28:29 -0500
Message-ID: <7897.1572964109@sss.pgh.pa.us>
Precedence: bulk

Stephen Frost <sfrost@snowman.net> writes:
> * Peter Eisentraut (peter.eisentraut@2ndquadrant.com) wrote:
>> On 2019-11-04 15:53, Alvaro Herrera wrote:
>>>> No, I'm not proposing a full revert.  But there's certainly room to
>>>> consider reverting the part that says you*must*  write "-f -" to get
>>>> output to stdout.

>>> I don't think this will buy us anything, if we get past branches updated
>>> promptly.

>> I think we should revert the part that requires using -f - at least for
>> PG12.

> Absolutely not.  This argument could be made, with a great deal more
> justification, against the changes to remove recovery.conf, and I'm sure
> quite a few other changes that we've made between major versions over
> the years, but to do so would be to hamstring our ability to make
> progress and to improve PG.

In this case, not in the least: we would simply be imposing the sort
of *orderly* feature introduction that I thought was the plan from
the very beginning [1].  That is, first make "-f -" available, and
make it required only in some later version.  If we'd back-patched
the optional feature back in April, it might've been okay to require
it in v12, but we failed to provide any transition period.

I'm in favor of making v12 act like the older branches now do, and
requiring "-f -" only as of v13.  Yeah, the transition will be a
little slower, but this feature is not of such huge value that it
really justifies breaking scripts with zero notice.

			regards, tom lane

[1] https://www.postgresql.org/message-id/24868.1550106683%40sss.pgh.pa.us