From: | Alvaro Herrera <alvherre(at)commandprompt(dot)com> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | Khee Chin <kheechin(at)gmail(dot)com>, pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: reloptions with a "namespace" |
Date: | 2009-04-03 20:40:41 |
Message-ID: | 20090403204041.GO23023@alvh.no-ip.org |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Tom Lane escribió:
> Alvaro Herrera <alvherre(at)commandprompt(dot)com> writes:
> > Tom Lane escribi:
> >> Surely this will break other things. I find myself wondering why you
> >> invented ReloptElem at all, instead of adding a field to DefElem.
>
> > I had to, precisely because it messes up other uses of DefElem ...
>
> > For example, the grammar would allow
> > CREATE FUNCTION ... WITH something.name = value
> > which we certainly don't want.
>
> Well, you could still have separate productions that did or didn't allow
> qualified names there (or perhaps better, have the code in
> functioncmds.c reject qualified names). I think the use of two different
> node types is going to result in duplicate coding and/or bugs deeper in
> the system, however.
I think what drove me away from that (which I certainly considered at
some point) was the existance of OptionDefElem. Maybe it would work to
make RelOptElem similar to that, i.e. have a char *namespace and a
DefElem?
--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support
From | Date | Subject | |
---|---|---|---|
Next Message | Alvaro Herrera | 2009-04-03 20:43:40 | Re: reloptions with a "namespace" |
Previous Message | Tom Lane | 2009-04-03 20:29:25 | Re: a few crazy ideas about hash joins |