Re: [ADMIN] Schema comparisons - Mailing list pgsql-hackers

From Richard Huxton
Subject Re: [ADMIN] Schema comparisons
Date
Msg-id 200402291408.07797.dev@archonet.com
Whole thread Raw
In response to Re: [ADMIN] Schema comparisons  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
On Sunday 29 February 2004 02:01, Tom Lane wrote:
> Richard Huxton <dev@archonet.com> writes:
> > I've been looking at storing $REVISION$ in comments for each object, so
> > my install scripts can halt if there is a problem. Not wanting to use my
> > only comment slot for this I was thinking about an extension to the
> > COMMENT ON statement:
> > COMMENT ON TABLE foo IS 'This is where I stroe my foos.';
> > COMMENT ON TABLE foo SECTION 'default' IS 'I meant store my foos.';
> > COMMENT ON TABLE foo SECTION 'revision' IS '1.19';
> > COMMENT ON TABLE foo SECTION 'bar' IS 'baz';
>
> This seems a little, um, specialized.  Why don't you just keep the info
> in a user-defined table?

For the same reasons you don't store existing comments in a user-defined 
table:
1. It's convenient to have a standard (across providers) place for them.
2. It's meta-data, not data.
3. It gets dumped along with my table.

If it's just a case of "looks like a waste of time" then I might well waste my 
time and do it. On the other hand, if it's a case of "unnecessary 
complication - don't want it in the code" then I'll not bother.
--  Richard Huxton Archonet Ltd


pgsql-hackers by date:

Previous
From: Andrew Dunstan
Date:
Subject: Re: [PATCHES] log_line_info
Next
From: Neil Conway
Date:
Subject: Re: Avoid MVCC using exclusive lock possible?