Thread: Misplaced BKI entries in pg_amproc.h

Misplaced BKI entries in pg_amproc.h

From
Antonin Houska
Date:
While checking something, I noticed that opfamilies 3626, 3683, 3901 
(all btree AM), 3903 (hash) and 3919 (gist) are all defined in the 
section marked as "gin".

(I'm not sure if it helps to deliver a patch - it may be easier for the 
committer to move the items himself than to check if the diff is correct)

// Antonin Houska (Tony)



Re: Misplaced BKI entries in pg_amproc.h

From
Bruce Momjian
Date:
On Thu, Aug  1, 2013 at 10:14:43AM +0200, Antonin Houska wrote:
> While checking something, I noticed that opfamilies 3626, 3683, 3901
> (all btree AM), 3903 (hash) and 3919 (gist) are all defined in the
> section marked as "gin".
>
> (I'm not sure if it helps to deliver a patch - it may be easier for
> the committer to move the items himself than to check if the diff is
> correct)

You are absolutely correct.  I checked your report by comparing
src/include/catalog/pg_amproc.h against this query:

    SELECT pg_opfamily.oid, pg_opfamily.*, pg_am.amname
    FROM pg_opfamily join pg_am ON (pg_opfamily.opfmethod = pg_am.oid)
    ORDER BY 2, 1;

and you are right that the ones you mentioned were in the wrong section.
I also reordered a few so the entries in each section were in increasing
order.

Attached patch applied.

--
  Bruce Momjian  <bruce@momjian.us>        http://momjian.us
  EnterpriseDB                             http://enterprisedb.com

  + Everyone has their own god. +

Attachment