Re: 7.3 gotchas for applications and client libraries - Mailing list pgsql-hackers

From Tom Lane
Subject Re: 7.3 gotchas for applications and client libraries
Date
Msg-id 5032.1031058971@sss.pgh.pa.us
Whole thread Raw
In response to 7.3 gotchas for applications and client libraries  (Lee Kindness <lkindness@csl.co.uk>)
Responses Re: 7.3 gotchas for applications and client libraries  (Lee Kindness <lkindness@csl.co.uk>)
List pgsql-hackers
Lee Kindness <lkindness@csl.co.uk> writes:
>  CREATE OR REPLACE FUNCTION column_exists(NAME, NAME) RETURNS BOOLEAN AS '

>  CREATE OR REPLACE FUNCTION table_exists(NAME) RETURNS BOOLEAN AS '

> Obviously these need attention when our application targets 7.3 (and
> thanks for the heads-up), but all changes are localised.

They are?  What will your policy be about schema names --- won't you
have to touch every caller to add a schema name parameter?

I'm not averse to trying to push logic over to the backend, but I think
the space of application requirements is wide enough that designing
general-purpose functions will be quite difficult.
        regards, tom lane


pgsql-hackers by date:

Previous
From: "Nigel J. Andrews"
Date:
Subject: Re: Memory management question
Next
From: leozc@cse.unsw.edu.au (Zhicong Leo Liang)
Date:
Subject: serial type as foreign key referential integrity violation