Re: PERFORM effects FOUND patch (Was: [GENERAL] I must be blind...)

Поиск
Список
Период
Сортировка
От Tom Lane
Тема Re: PERFORM effects FOUND patch (Was: [GENERAL] I must be blind...)
Дата
Msg-id 14572.1024158496@sss.pgh.pa.us
обсуждение исходный текст
Ответ на PERFORM effects FOUND patch (Was: [GENERAL] I must be blind...)  ("Nigel J. Andrews" <nandrews@investsystems.co.uk>)
Список pgsql-hackers
"Nigel J. Andrews" <nandrews@investsystems.co.uk> writes:
> +         if (estate->eval_processed != 0)
> +             exec_set_found(estate, true);

To be actually useful the command would have to set FOUND to either
true or false depending on whether it computed a row or not.  So the
correct patch would be more like

+        exec_set_found(estate, (estate->eval_processed != 0));

Also, changing the parameter to exec_run_select as you did is wrong.
A multi-row query should be allowed to run to completion, I'd think.

As for whether to apply it or not --- the change seems reasonable if we
were working in a vacuum.  But I don't believe we invented PERFORM out
of whole cloth; surely there are other systems that we need to consider
compatibility with.

            regards, tom lane

В списке pgsql-hackers по дате отправления:

Предыдущее
От: Tom Lane
Дата:
Сообщение: Re: Queries using rules show no rows modified?
Следующее
От: Bruce Momjian
Дата:
Сообщение: Re: Integrating libpqxx