Обсуждение: psql 15beta1 does not print notices on the console until transaction completes

Hi all,

I notice this change in behaviour with psql in 15beta1 when testing an existing codebase.

I didn't see any mention of this change in the release notes and it surprised me.

Using this test function:

    create or replace function test_notice() returns void as
    $$
    begin
        raise notice 'hello';
        perform pg_sleep(10);
    end; $$ language plpgsql;

In psql 15beta1, the "hello" message only appears on the console when the transaction completes.

in psql 14.3, it appears immediately as I would have expected.

Is there a way to change psql behaviour to display notices immediately as in versions < 15?

Best regards,

Alastair

On 5/29/22 13:11, Alastair McKinley wrote:
> Hi all,
> 
> I notice this change in behaviour with psql in 15beta1 when testing an existing codebase.
> 
> I didn't see any mention of this change in the release notes and it surprised me.
> 
> Using this test function:
> 
>      create or replace function test_notice() returns void as
>      $$
>      begin
>          raise notice 'hello';
>          perform pg_sleep(10);
>      end; $$ language plpgsql;
> 
> In psql 15beta1, the "hello" message only appears on the console when the transaction completes.

I am not seeing that.

Can you provide more information about how you are running test_notice()?

> 
> in psql 14.3, it appears immediately as I would have expected.
> 
> Is there a way to change psql behaviour to display notices immediately as in versions < 15?
> 
> Best regards,
> 
> Alastair
> 


-- 
Adrian Klaver
adrian.klaver@aklaver.com

On 5/29/22 13:47, Adrian Klaver wrote:
> On 5/29/22 13:11, Alastair McKinley wrote:
>> Hi all,

>> In psql 15beta1, the "hello" message only appears on the console when 
>> the transaction completes.
> 
> I am not seeing that.

I take that back, I was using psql 14.3 to connect to the 15 instance. 
When I changed to psql 15 I saw the same thing.

> 

>> Best regards,
>>
>> Alastair
>>
> 
> 


-- 
Adrian Klaver
adrian.klaver@aklaver.com

On 5/29/22 13:59, Alastair McKinley wrote:
>  >
>  > From: Adrian Klaver <adrian.klaver@aklaver.com>
>  > Sent: 29 May 2022 21:47To: Alastair McKinley 

> Hi Adrian,
> 
> I am running the function "select test_notice();" from the psql console 
> with psql/server versions 15beta1.
> 
> In psql 15beta1, the notice appears only after the function 
> returns/transaction completes.
> 
> If I execute the same function from psql 14.3 and server 15beta1, the 
> notice appears on the psql console immediately as expected, before the 
> pg_sleep() completes.
> 
> It's reproducible for me, I just double checked it.

Yeah, I see the same thing  after using the new psql.

> 
> I think Tom was able to reproduce this by the sounds of his response?

I have not received that post yet. I do see it in the archives. I had 
already tried:

\set SHOW_ALL_RESULTS off

which is supposed to restore to previous behavior, but it did not lead 
to the 14.3 result.

> 
> Best regards,
> 
> Alastair
> 

-- 
Adrian Klaver
adrian.klaver@aklaver.com

Adrian Klaver <adrian.klaver@aklaver.com> writes:
> On 5/29/22 13:59, Alastair McKinley wrote:
>> I think Tom was able to reproduce this by the sounds of his response?

> I have not received that post yet. I do see it in the archives.

I re-addressed it to pgsql-bugs, maybe you are not subscribed to that?

            regards, tom lane

On 5/29/22 15:03, Tom Lane wrote:
> Adrian Klaver <adrian.klaver@aklaver.com> writes:
>> On 5/29/22 13:59, Alastair McKinley wrote:
>>> I think Tom was able to reproduce this by the sounds of his response?
> 
>> I have not received that post yet. I do see it in the archives.
> 
> I re-addressed it to pgsql-bugs, maybe you are not subscribed to that?

I am not, still I see it here:

https://www.postgresql.org/list/pgsql-general/since/202205290000/

And when I click on the message:

https://www.postgresql.org/message-id/3662994.1653856025%40sss.pgh.pa.us

Lists:     pgsql-bugs pgsql-general

I will admit that the mailing list software is a mystery to me, so it is 
entirely possible I am misreading the above.

> 
>             regards, tom lane


-- 
Adrian Klaver
adrian.klaver@aklaver.com

On 5/29/22 18:00, Adrian Klaver wrote:
> On 5/29/22 15:03, Tom Lane wrote:
>> Adrian Klaver <adrian.klaver@aklaver.com> writes:
>>> On 5/29/22 13:59, Alastair McKinley wrote:
>>>> I think Tom was able to reproduce this by the sounds of his response?
>>
>>> I have not received that post yet. I do see it in the archives.
>>
>> I re-addressed it to pgsql-bugs, maybe you are not subscribed to that?
>
> I am not, still I see it here:
>
> https://www.postgresql.org/list/pgsql-general/since/202205290000/
>
> And when I click on the message:
>
> https://www.postgresql.org/message-id/3662994.1653856025%40sss.pgh.pa.us
>
> Lists:     pgsql-bugs pgsql-general
>
> I will admit that the mailing list software is a mystery to me, so it 
> is entirely possible I am misreading the above.
>
FWIW, I didn't see Tom's first (bugs) post either.  Nor am I subscribed 
to bugs