Date: Thu, 1 Sep 2022 16:31:20 -0700
From: Nathan Bossart <nathandbossart@gmail.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Robert Haas <robertmhaas@gmail.com>,
	Peter Eisentraut <peter.eisentraut@enterprisedb.com>,
	pgsql-hackers <pgsql-hackers@postgresql.org>
Subject: Re: postgres_fdw hint messages
Message-ID: <20220901233120.GA934376@nathanxps13>
References: <b1f9f399-3a1a-b554-283f-4ae7f34608e2@enterprisedb.com>
 <CA+TgmoY+BuSZm-Fbxx5zC6SuML=2F9YvaYuzVAJ=XyE65BnchQ@mail.gmail.com>
 <1929340.1661531738@sss.pgh.pa.us>
 <20220901223128.GA932597@nathanxps13>
 <789102.1662073729@sss.pgh.pa.us>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <789102.1662073729@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/20220901233120.GA934376%40nathanxps13>
Precedence: bulk

On Thu, Sep 01, 2022 at 07:08:49PM -0400, Tom Lane wrote:
> (1) there probably needs to be some threshold of closeness, so we don't
> offer "foobar" when the user wrote "huh"

Agreed.

> (2) there are several places doing this now, and there will no doubt
> be more later, so we need to try to abstract the logic so it can be
> shared.

Will do.

I'm also considering checking whether the user-provided string is longer
than MAX_LEVENSHTEIN_STRLEN so that we can avoid the ERROR from
varstr_levenshtein().  Or perhaps varstr_levenshtein() could indicate that
the string is too long without ERROR-ing (e.g., by returning -1).  If the
user-provided string is too long, we'd just omit the hint.

-- 
Nathan Bossart
Amazon Web Services: https://aws.amazon.com