MIME-Version: 1.0
References: <ahm_4eOKkkKJ3Gds@nathan>
 <CAOYmi+msZYr=NhBmf-Tg3h3ZHoQjMzp-50ByFqo2A8XDKd-eHQ@mail.gmail.com>
 <CAGECzQQCKSogfz=Dqkf-+1GA13yATMu34CMU9Yqo13KXgmkW8Q@mail.gmail.com>
 <aiNHmeERanDZViYv@nathan>
In-Reply-To: <aiNHmeERanDZViYv@nathan>
From: Jelte Fennema-Nio <postgres@jeltef.nl>
Date: Sat, 6 Jun 2026 00:24:48 +0200
Message-ID: 
 <CAGECzQRwr3yFmh1Hf9wyyYR0DjuSi0ZCiWHvYyY4pUp_ALF1bg@mail.gmail.com>
Subject: Re: alert clients when prepared statements are deallocated
To: Nathan Bossart <nathandbossart@gmail.com>
Cc: Jacob Champion <jacob.champion@enterprisedb.com>,
	PostgreSQL-development <pgsql-hackers@postgresql.org>
Content-Type: multipart/alternative; boundary="000000000000a5dbfa06538921a1"
Archived-At: 
 <https://www.postgresql.org/message-id/CAGECzQRwr3yFmh1Hf9wyyYR0DjuSi0ZCiWHvYyY4pUp_ALF1bg%40mail.gmail.com>
Precedence: bulk

--000000000000a5dbfa06538921a1
Content-Type: text/plain; charset="UTF-8"

On Sat, Jun 6, 2026, 00:03 Nathan Bossart <nathandbossart@gmail.com> wrote:

> Do we need to guard who can create protocol-level statements?  And if so,
> how would we do that?
>

I think the only thing we'd need to guard against is a libpq user messing
with libpq its own internal protocol-level namespace, let's call that
"pqnamespace" for now. We'd add a new version of PQsendPrepare that would
take a namespace, but we'd make that return an error if a user of libpq
passed pqnamespace as the namespace argument. Similarly for the new
PQsendClosePrepared variant that takes a namespace. That way we'd reserve
that namespace.

>

--000000000000a5dbfa06538921a1
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"auto"><div><div class=3D"gmail_quote gmail_quote_container"><di=
v dir=3D"ltr" class=3D"gmail_attr">On Sat, Jun 6, 2026, 00:03 Nathan Bossar=
t &lt;<a href=3D"mailto:nathandbossart@gmail.com">nathandbossart@gmail.com<=
/a>&gt; wrote:</div><blockquote class=3D"gmail_quote" style=3D"margin:0 0 0=
 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Do we need to guard who can create protocol-level statements?=C2=A0 And if =
so,<br>
how would we do that?<br></blockquote></div></div><div dir=3D"auto"><br></d=
iv><div dir=3D"auto">I think the only thing we&#39;d need to guard against =
is a libpq user messing with libpq its own internal protocol-level namespac=
e, let&#39;s call that &quot;pqnamespace&quot; for now. We&#39;d add a new =
version of PQsendPrepare that would take a namespace, but we&#39;d make tha=
t return an error if a user of libpq passed pqnamespace as the namespace ar=
gument. Similarly for the new PQsendClosePrepared variant that takes a name=
space. That way we&#39;d reserve that namespace.</div><div dir=3D"auto"><di=
v class=3D"gmail_quote gmail_quote_container"><blockquote class=3D"gmail_qu=
ote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex=
">
</blockquote></div></div></div>

--000000000000a5dbfa06538921a1--