Re: Re: [Pg-migrator-general] Composite types break pg_migratedtables - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Re: [Pg-migrator-general] Composite types break pg_migratedtables
Date
Msg-id 3615.1249569418@sss.pgh.pa.us
Whole thread Raw
In response to Re: Re: [Pg-migrator-general] Composite types break pg_migratedtables  (Andrew Dunstan <andrew@dunslane.net>)
List pgsql-hackers
Andrew Dunstan <andrew@dunslane.net> writes:
> Kevin Grittner wrote:
>> Any idea whether domains are an issue?

> I don't believe that they are an issue. The issue arises only when a 
> catalog oid is used in the on-disk representation of a type. AFAIK the 
> on-disk representation of a domain is the same as its base type.

Arrays of domains would be a problem, if we had 'em, which we don't...

Also, as Peter already noted, arrays of built-in types are not really a
problem because the OID won't have changed since 8.3.  It's only arrays
of types created post-initdb that are risk factors.
        regards, tom lane


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Adding error message "source"
Next
From: Andrew Dunstan
Date:
Subject: Re: Re: [Pg-migrator-general] Composite types break pg_migrated tables