Message-ID: <011d01bf7357$4c52b160$2dd260d1@arcamax.com>
From: "Bryan White" <bryan@arcamax.com>
To: <pgsql-interfaces@postgreSQL.org>
References: <XFMail.000209141919.orsingerc@epg-gw1.lewis.army.mil>
Subject: Re: [INTERFACES] The persistance of C functions
Date: Wed, 9 Feb 2000 18:42:16 -0500
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit

> On 09-Feb-00 Bryan White wrote:
> > If I use CREATE FUNCTION for a C function in a .so file and then use the
> > function and then change and recompile the function, what steps are
needed
> > to see the change?
> >
> > As I see it the options are:
> > A: do nothing, the function is reloaded on every invocation.
> > B: Reopen the connection to the backend
> > C: Restart the postmaster
> >
> > I suspect B is correct but I would like to hear someone confirm it.
>
>         I don't know about A or B, but C definitely works :-). If it's
> practical, I sometimes use DROP FUNCTION/CREATE FUNCTION, but I don't
> think this is practical when you're using this function (or set of
> functions) to implement a new data type.

Thanks, I have come to the conclusion that B is sufficient based on trial
and error.