Re: pg_upgrade - Mailing list pgsql-general

From Brian Hirt
Subject Re: pg_upgrade
Date
Msg-id C4222AE3-7E70-4008-9882-ECE71615E0D9@me.com
Whole thread Raw
In response to Re: pg_upgrade  (Bruce Momjian <bruce@momjian.us>)
Responses Re: pg_upgrade  (Bruce Momjian <bruce@momjian.us>)
List pgsql-general
Bruce,

The applied patch has the same behavior on i686 Ubuntu 10.04.   It looks like atol() is just a macro for strtol() in
stdio.h.  I think you want strtoul() instead of strtol() 

when i change str2uint() to use strtoul() pg_upgrade completes without a problem (I still haven't tested the upgrade
database,but I expect that will be just fine). 

I think it's pretty uncommon for the OID to be that big which is why nobody stumbled onto this.   This particular
installationhas pretty much been reloading development databases non stop for the last year.  Also, people tend to
initdba lot when testing and doing development which will keep resetting the oid low. 

Thanks for getting this one fixed

--brian

On Sep 28, 2010, at 3:49 PM, Bruce Momjian wrote:
> I have applied the attached patch to HEAD and 9.0.X.  Odd I had never
> received a bug report about this before.  Good thing it didn't silently
> fail, but it is designed to be very picky.
>

pgsql-general by date:

Previous
From: Bartlomiej Korupczynski
Date:
Subject: How to handle results with column names clash
Next
From: Eric McKeeth
Date:
Subject: Re: Exclusion constraint issue