Re: No, pg_size_pretty(numeric) was not such a hot idea

On Tue, Jun 5, 2012 at 1:01 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Jim Nasby <jim@nasby.net> writes:
>> On 5/27/12 2:54 PM, Euler Taveira wrote:
>>> On 27-05-2012 10:45, Fujii Masao wrote:
>>>> OK, let me propose another approach: add pg_size_pretty(int).
>
>>> I wouldn't like to add another function but if it solves both problems... +1.
>
>> FWIW, I would argue that the case of pg_size_pretty(8*1024*1024) is
>> pretty contrived...
>
> Yeah, possibly.  In any case, I don't think we're making either of these
> changes in 9.2, because the time for forcing initdbs is past.  It would
> only be realistic to think about changing pg_size_pretty() if we come
> across some other, much more compelling reason to force a system catalog
> contents change.
>
> Assuming that's how 9.2 ships, we might as well wait to see if there
> are any real complaints from the field before we decide whether any
> changing is needed.

We could add it to the catalog without forcing an initdb. That way
anybody who installed on the release (or beta3+) would get the
function, and not those who started on the beta (unless they created
it manually). That does leave us in a position where people have
different versions of the schema with the same identifier though, so
that may not be the best idea.

If we're just leaving it, should we take it off the open items list,
or leave it in there "in case something else shows up"?

--
 Magnus Hagander
 Me: http://www.hagander.net/
 Work: http://www.redpill-linpro.com/