Re: RfD: more powerful "any" types

Поиск
Список
Период
Сортировка
От Tom Lane
Тема Re: RfD: more powerful "any" types
Дата
Msg-id 13655.1252430928@sss.pgh.pa.us
обсуждение исходный текст
Ответ на Re: RfD: more powerful "any" types  (Alvaro Herrera <alvherre@commandprompt.com>)
Список pgsql-hackers
Alvaro Herrera <alvherre@commandprompt.com> writes:
> David E. Wheeler wrote:
>> Why can't anyelement be freed from this constraint?

> Because it would break other uses of it, I think.

Specifically, what are you going to do with something like
make_array(anyelement, anyelement) returns anyarray

There's no principled way to determine what anyarray means if the
arguments are not the same type.

I have no objection to adding some other pseudotype with different
behavior, but breaking anyelement is not the path.
        regards, tom lane


В списке pgsql-hackers по дате отправления:

Предыдущее
От: "David E. Wheeler"
Дата:
Сообщение: Re: RfD: more powerful "any" types
Следующее
От: Tom Lane
Дата:
Сообщение: Re: RfD: more powerful "any" types