Re: CLASSOID patch - Mailing list pgsql-hackers

From Chris Bitmead
Subject Re: CLASSOID patch
Date
Msg-id 3956CD08.8124E5EB@nimrod.itg.telecom.com.au
Whole thread Raw
In response to Re: CLASSOID patch  (Peter Eisentraut <peter_e@gmx.net>)
Responses Re: CLASSOID patch  (M.Feldtmann@t-online.de (Marten Feldtmann))
List pgsql-hackers
Peter Eisentraut wrote:
> 
> Chris Bitmead writes:
> 
> > Attached is a first attempt at implementing the classoid feature.
> 
> I'm wondering what other people think about the naming. Firstly, it's my
> feeling that TABLEOID would be more in line with the general conventions.

I was thinking this myself today. Mainly because I wonder if in the
future there may be support for more than one table implementing a
particular class type. On the other hand the oid is a reference to the
pg_class table. Maybe pg_class should be renamed pg_table? Anyway, my
current thinking is that tableoid is better.

The general naming conventions in postgres are a bit disturbing. Some
places refer to classes, some to tables, some to relations. One day it
should all be reconciled :-).

> Secondly, maybe we ought to make the name less susceptible to collision by
> choosing a something like _CLASSOID (or whatever).

Only if oid becomes _oid and ctid becomes _ctid. I don't think it's
worth it myself.

> > It works!
> 
> Great! :)


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Maximum len of data fit into the tuple
Next
From: Tom Lane
Date:
Subject: Re: Re: Call for port testing on fmgr changes -- Results!