Re: Fixing busted citext function declarations - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Fixing busted citext function declarations
Date
Msg-id 2090.1431041516@sss.pgh.pa.us
Whole thread Raw
In response to Re: Fixing busted citext function declarations  ("David G. Johnston" <david.g.johnston@gmail.com>)
List pgsql-hackers
"David G. Johnston" <david.g.johnston@gmail.com> writes:
> On Thu, May 7, 2015 at 4:19 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> pg_upgrade is okay in any case because it dumps and reloads the current
>> extension's components.  Doesn't matter whether there's another version
>> that is not compatible.

> ​For clarity - which one is "current" in this context?

> 1. The existing database's (previous extension version)
> 2. The target database's (current default extension version in the new
> PostgreSQL version)

> ​The answer has to be #2 since the version in the existing database no
> longer exists in the new PostgreSQL version.

You're mistaken.  pg_dump --binary_upgrade does not care whether the
target database thinks that version exists or not.  (It does care that
there's a compatible shared-library object, but that's not at issue
in this case.)
        regards, tom lane



pgsql-hackers by date:

Previous
From: "David G. Johnston"
Date:
Subject: Re: Fixing busted citext function declarations
Next
From: Michael Paquier
Date:
Subject: Re: commitfest app bug/feature