Re: [HACKERS] Scripts again
От | Peter Eisentraut |
---|---|
Тема | Re: [HACKERS] Scripts again |
Дата | |
Msg-id | Pine.GSO.4.02A.9910141258330.29358-100000@Ekorre.DoCS.UU.SE обсуждение исходный текст |
Ответ на | Re: [HACKERS] Scripts again (wieck@debis.com (Jan Wieck)) |
Ответы |
Re: [HACKERS] Scripts again
|
Список | pgsql-hackers |
On Wed, 13 Oct 1999, Jan Wieck wrote: > Make the whole thing configurable and anyone should be happy. > > --pg_admin_script_prefix={pg_|pg|*empty*|*whatever_you_prefer*} > --pg_admin_script_install={yes|no} Way initially I was suggesting--enable-scripts=old|new|both|none (default new) but Bruce found *that* too complicated. I can see your point here, accessing different db installations, but I think that is a highly specialized case (and you should be using psql anyway, but that seems to be a culture issue). But the scripts have no concept of default ports etc., that's in libpq. So you can use pg_createdb for whatever your default install is, and pg_createdb -p foo for your alternate installation. Or you could alias this or something. (Btw., anyone else think a /etc/services entry is better than a hardwired default port, at least on the libpq side of things? Of course, I'm not sure about Windows here.) > > It's not a joke. Someone might want to have his user account > to have access to his production and test DB at the same > time. So he could setup his PATH to both installations bin > directories and configure the test DB to use a different > default PGPORT and different script prefixes. Then > pg_createdb would contact another postmaster than > devel_createdb would do. Well, the installed binaries like > psql would need some configurable prefix too then. > > > Jan > -- Peter Eisentraut Sernanders vaeg 10:115 peter_e@gmx.net 75262 Uppsala http://yi.org/peter-e/ Sweden
В списке pgsql-hackers по дате отправления: