Re: SRFs ExecMakeTableFunctionResult

Поиск
Список
Период
Сортировка
От Tom Lane
Тема Re: SRFs ExecMakeTableFunctionResult
Дата
Msg-id 14433.1092061555@sss.pgh.pa.us
обсуждение исходный текст
Ответ на SRFs ExecMakeTableFunctionResult  (James William Pye <flaw@rhid.com>)
Ответы Re: SRFs ExecMakeTableFunctionResult  (James William Pye <flaw@rhid.com>)
Список pgsql-hackers
James William Pye <flaw@rhid.com> writes:
> While I was finishing up SRF support in PL/Py, I noticed that when VPC is the
> selected mode for a table function, ExecMakeTableFunctionResult will set
> rsinfo.isDone to ExprSingleResult each time it loops to fetch another value
> (when a direct_function_call). This makes the VPC-SRF author set isDone to
> ExprMultipleResult on _every_ call while returning values, as it will break
> out if rsinfo.isDone !=3D ExprMultipleResult.

> Is this the desired behavior?

Seems reasonable to me.  A SRF function really ought to explicitly set
isDone on every call anyway.
        regards, tom lane


В списке pgsql-hackers по дате отправления:

Предыдущее
От: Tom Lane
Дата:
Сообщение: Re: 8.0.0beta1 ... packaged for testing ...
Следующее
От: Tom Lane
Дата:
Сообщение: Re: Changing the type of timestamp columns