Re: LISTEN considered dangerous
От | Ian Harding |
---|---|
Тема | Re: LISTEN considered dangerous |
Дата | |
Msg-id | 725602300608020646gb908d7et1e91cfe24aab86f@mail.gmail.com обсуждение исходный текст |
Ответ на | Re: LISTEN considered dangerous (Flemming Frandsen <ff@partyticket.net>) |
Список | pgsql-general |
On 8/2/06, Flemming Frandsen <ff@partyticket.net> wrote: > Ian Harding wrote: > > NOTIFY interacts with SQL transactions in some important ways. > > Firstly, if a NOTIFY is executed inside a transaction, the notify > > events are not delivered until and unless the transaction is > > committed. This is appropriate, since if the transaction is aborted, > > all the commands within it have had no effect, including NOTIFY. But > > it can be disconcerting if one is expecting the notification events to > > be delivered immediately. > > Yes, that's very nice, but it doesn't have *anything* to do with what I > posted about. > Quite true, but it does indicate, to me at least, the fact that this is a SQL command and doesn't take effect until committed. From what I read in the docs, I would expect the NOTIFY signals to be like phone calls, if your phone's not plugged in (LISTEN not committed) you miss the call. That's the way it works apparently. > I'm bothered by listen listening from the end of the transaction in > stead of the start of the transaction. > What seems to be needed is an answering service that will record your NOTIFY events, in case you decide to plug in the phone and retrieve them. - Ian
В списке pgsql-general по дате отправления: