From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Michael Meskes <meskes(at)postgresql(dot)org> |
Cc: | PostgreSQL Hacker <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: type OIDs |
Date: | 2000-09-15 14:23:01 |
Message-ID: | 20177.969027781@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Michael Meskes <meskes(at)postgresql(dot)org> writes:
> Can I safely assume that the OID of the standard data types remain the same
> for future releases? And of course that they are the same for every
> installation?
They are fixed in any one version, and really are not very likely to
change across versions either. But I suppose it could happen.
> I've been send a patch to speed up ecpg significantly by not looking up
> datatypes everytime. As it is written right now it works by har coding some
> types. I wonder if this will create problems.
Exactly how "hard coded" do you mean? If you #include "catalog/pg_types.h"
and use the OID #defines therein, you're not doing any worse than a lot
of places in the backend. At worst you'd create a cross-major-version
incompatibility for ecpg.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2000-09-15 14:24:30 | Re: Cannot compile |
Previous Message | Philip Warner | 2000-09-15 12:07:11 | Re: current is broken |