On Mon, Apr 25, 2011 at 7:03 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Robert Haas <robertmhaas@gmail.com> writes:
>> The sequence of steps that he posted wasn't actually right.
>
> Yes, I did read the thread.
OK.
>> The patch seems trivially correct,
>> since this is obviously schizophrenic:
>
>> ereport(ERROR,
>> (errcode(ERRCODE_WRONG_OBJECT_TYPE),
>> errmsg("\"%s\" is not a foreign table",
>> ! RelationGetRelationName(targetrelation)),
>> ! errhint("Use ALTER FOREIGN TABLE instead.")));
>
> Well, it's a pretty obvious global-search-and-replace error, but I fail
> to see the advantage of just deleting the hint. I was thinking
>
> - errhint("Use ALTER FOREIGN TABLE instead.")));
> + errhint("Use ALTER TABLE instead.")));
>
> As per the other thread today, this advice would usually be correct,
> so I think that not offering any advice at all would be a step down from
> that.
Well, currently ALTER TABLE will work even if the argument is a view
or sequence, but I view that as a backwards-compatibility kludge we
should be looking to move away from, not something we want to further
bake in. However, I'm out of time to bikeshed on this issue, so fix
it however you like.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company