Re: Incorrect processing of CREATE TRANSFORM with DDL deparding

On Fri, Jun 5, 2015 at 6:24 AM, Alvaro Herrera <alvherre@2ndquadrant.com> wrote:
>
> Michael Paquier wrote:
>
> > And I actually bumped into a bug with this fake test case:
> > DROP TRANSFORM FOR int LANGUAGE SQL;
> > + ERROR:  requested object address for unsupported object class 32: text
> > result "unrecognized object 3576 16551 0"
> > What happens is that the handling for transforms in getObjectIdentityParts
> > is missing. This is fixed as well in the attached.
>
> Oh, interesting.  We need one or two extra lines in the
> object_address.sql test as well, I guess.


Ah, so that's where it is... Attached is an updated patch with this
stuff. I noticed a couple of code paths where OCLASS_TRANSFORM was
missing as well.
--
Michael