Re: Make query cancellation keys longer

From: Heikki Linnakangas <hlinnaka(at)iki(dot)fi>
To: Robert Haas <robertmhaas(at)gmail(dot)com>
Cc: Jelte Fennema-Nio <postgres(at)jeltef(dot)nl>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>, Thomas Munro <thomas(dot)munro(at)gmail(dot)com>
Subject: Re: Make query cancellation keys longer
Date: 2024-08-15 22:07:25
Message-ID: 08fee89e-0db2-4545-b00a-ce3436ca5075@iki.fi
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 15/08/2024 23:20, Robert Haas wrote:
> On Thu, Aug 15, 2024 at 1:13 PM Heikki Linnakangas <hlinnaka(at)iki(dot)fi> wrote:
>> Added a "protocol_version" libpq option for that. It defaults to "auto",
>> but you can set it to "3.1" or "3.0" to force the version. It makes it
>> easier to test that the backwards-compatibility works, too.
>
> Over on the "Add new protocol message to change GUCs for usage with
> future protocol-only GUCs" there is a lot of relevant discussion about
> how bumping the protocol version should work. This thread shouldn't
> ignore all that discussion. Just to take one example, Jelte wants to
> bump the protocol version to 3.2, not 3.1, for some reasons that are
> in the commit message for the relevant patch over there.

Ok, I've read through that thread now, and opined there too. One
difference is with libpq option name: My patch adds "protocol_version",
while Jelte proposes "max_protocol_version". I don't have strong
opinions on that. I hope the ecosystem catches up to support
NegotiateProtocolVersion quickly, so that only few people will need to
set this option. In particular, I hope that there will never be need to
use "max_protocol_version=3.2", because by the time we introduce version
3.3, all the connection poolers that support 3.2 will also implement
NegotiateProtocolVersion.

--
Heikki Linnakangas
Neon (https://neon.tech)

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Peter Geoghegan 2024-08-15 22:34:16 Re: Showing primitive index scan count in EXPLAIN ANALYZE (for skip scan and SAOP scans)
Previous Message Heikki Linnakangas 2024-08-15 22:03:57 Re: Add new protocol message to change GUCs for usage with future protocol-only GUCs