Re: BUG #16258: exec_simple_query does not call ProcessCompletedNotifiesafter CommitTransactionCommand - Mailing list pgsql-bugs

From RekGRpth
Subject Re: BUG #16258: exec_simple_query does not call ProcessCompletedNotifiesafter CommitTransactionCommand
Date
Msg-id CAPgh2mKn8ZB_vsrjnS9JsYGDuKVa4nZ4D-2d6yiNK_pO67C8=Q@mail.gmail.com
Whole thread Raw
In response to Re: BUG #16258: exec_simple_query does not call ProcessCompletedNotifiesafter CommitTransactionCommand  (Jeff Janes <jeff.janes@gmail.com>)
Responses Re: BUG #16258: exec_simple_query does not call ProcessCompletedNotifiesafter CommitTransactionCommand
List pgsql-bugs
Thanks for the tip! I learned outer code of function exec_simple_query and I seen "idle" condition of calling ProcessCompletedNotifies! In my postgres job scheduler https://github.com/RekGRpth/pg_task I copied exec_simple_query and modified it to call ProcessCompletedNotifies after CommitTransactionCommand and it works!
with bst regrds, Rek>pth


сб, 15 февр. 2020 г. в 23:16, Jeff Janes <jeff.janes@gmail.com>:
On Sat, Feb 15, 2020 at 4:39 AM RekGRpth <rekgrpth@gmail.com> wrote:

And how about several transaction in new procedures?

That does seem like a legitimate gripe.

create or replace procedure foo() LANGUAGE plpgsql as $$
  begin 
  FOR i IN 1..100 LOOP
    perform pg_notify('channel','HI');
    perform pg_sleep(1); 
    commit;
  end loop;
end; $$;
CALL foo();

An actively polling listener gets all notification all at once at the end.  (I just use psql with a "LISTEN channel;" and the repeatedly execute ";" to get it to poll)

 Cheers,

Jeff

pgsql-bugs by date:

Previous
From: Tom Lane
Date:
Subject: Re: Possible memory leak in PostgreSQL 12.2 on truncate, insert or delete when type interval is involved
Next
From: "David G. Johnston"
Date:
Subject: Re: PSQL Help