Re: pg_dump ordering problem (rc4) - Mailing list pgsql-hackers

From Tom Lane
Subject Re: pg_dump ordering problem (rc4)
Date
Msg-id 14394.987267030@sss.pgh.pa.us
Whole thread Raw
In response to Re: pg_dump ordering problem (rc4)  (Philip Warner <pjw@rhyme.com.au>)
List pgsql-hackers
Philip Warner <pjw@rhyme.com.au> writes:
> I don't suppose we can change the pg_views view without an initdb?

No, not very readily.

Even assuming that we were willing to require dbadmins to run a script
during upgrade (which I wouldn't want to do unless forced to it), it's
not that easy to fix all the occurrences of a system view.  Remember
there will be a separate copy in each database, including template0
which you can't even get to.  On top of which, DROP VIEW/CREATE VIEW
wouldn't work because the view would then have the wrong OID, and would
look like a user-created object to pg_dump.  You'd have to manually
manipulate tuples in pg_class, pg_attribute, pg_rewrite, etc.

Kluging pg_dump is a *lot* cleaner.

I agree with the idea of adding the rule OID as a new column of pg_views
for 7.2, however.
        regards, tom lane


pgsql-hackers by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: Re: 7.1 RPMs
Next
From: Tom Lane
Date:
Subject: Re: Re: Possible explanation for readline configuration problems