Re: Postgres config file: autocommit = off
От | Rasmus Resen Amossen |
---|---|
Тема | Re: Postgres config file: autocommit = off |
Дата | |
Msg-id | Law14-F113A3MQiZij000005e8f@hotmail.com обсуждение исходный текст |
Ответ на | Postgres config file: autocommit = off ("Rasmus Resen Amossen" <rresena@hotmail.com>) |
Ответы |
Re: Postgres config file: autocommit = off
|
Список | pgsql-hackers |
>One of the reasons for taking autocommit control out of the backend and >pushing it up to the client level is exactly to make it feasible to take >these sorts of application-level considerations into account when >choosing the behavior. Ok, I can see some sense in that: Make the autocommit-behavior client dependent instead of system dependent. But that requires that all clients the user uses can handle this (is able to store a default behavior). I aggree, that clients should, as you write, overrule the system default behavior. But I (still) can't find an argument for, why the administrator should not have the oppotunity to set a default behavior for the whole system (not even in the archives). In this way postgres would be able to deal with clients that did not have support for setting the default behavior. Eventually a per user or per database default behavior could be usefull for the same resons. Bennefits: - Project managers can easier force programmers to use a specific database coding style. Fx.: I guess that if the PHP-interface should have a default value it should be given at the connect time. Programmers could easily forget to set "autocommit = off" here, thus allowing them self an eventually unwanted coding style. - Clinents which do not support setting an autocommit default behavior, can be used by setting the wanted behavior for the database system. Drawbacks: - ? (Enlighten me) _________________________________________________________________ Send s�de postkort til s�de mennesker http://www.msn.dk/postkort
В списке pgsql-hackers по дате отправления: