Subject: Re: Loose ends after CVE-2020-14350 (extension installation hazards)
To: Tom Lane <tgl@sss.pgh.pa.us>
References: <653936.1597431032@sss.pgh.pa.us>
 <5F36E107.50907@anastigmatix.net> <655710.1597433892@sss.pgh.pa.us>
Cc: pgsql-hackers@lists.postgresql.org
From: Chapman Flack <chap@anastigmatix.net>
Message-ID: <5F36F1C6.2080307@anastigmatix.net>
Date: Fri, 14 Aug 2020 16:19:18 -0400
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101
 Thunderbird/38.7.0
MIME-Version: 1.0
In-Reply-To: <655710.1597433892@sss.pgh.pa.us>
Content-Type: text/plain; charset=windows-1252
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 08/14/20 15:38, Tom Lane wrote:

> (3) If the SQL syntax is really just "WITH variable value [, ...]"
> then I'm afraid we're going to have a lot of parse-ambiguity problems
> with wedging full SET syntax into that.  The ability for the righthand

There is precedent in the SET command for having one general syntax
usable for any GUC, and specialized ones for a few 'special' GUCs
(search_path, client_encoding, timezone).

Maybe WITH could be done the same way, inventing some less thorny syntax
for the general case

   WITH (foo = bar, baz), (quux = 42), XMLBINARY BASE64, a AS (SELECT...)

and treating just the few like XMLBINARY that appear in the standard
as having equivalent specialized productions?

The only examples of the syntax in the standard that are coming to mind
right now are those I've seen in the SQL/XML part, but I feel like I have
seen others, as if the committee kind of likes their WITH local-setting-
of-something syntax, and additional examples may continue to appear.

Regards,
-Chap