Re: Fwd: Proposal: variant of regclass

On Sat, Mar 22, 2014 at 1:17 PM, Amit Kapila <amit.kapila16@gmail.com> wrote:
> On Wed, Feb 26, 2014 at 11:56 AM, Yugo Nagata <nagata@sraoss.co.jp> wrote:
>> Thanks for your a lot of comments. I revised the patch according to
>> comments from Robert Haas and Marti Raudsepp.
>
> I have started looking into this patch and below are my
> initial findings:
>
> 1. Dependency is not recorded when to_regclass is used,
> however it is recorded when ::regclass is used. Below is
> test to demonstrate this point. This change in behaviour is
> neither discussed nor mentioned in docs along with patch.

I think this is expected as per current design, because for
functions, it will create dependency on function (funcid), but
not on it's argument values. So I think for this behaviour, we might
need to mention in docs that to_regclass() and other newly added
functions will behave differently for creation of dependencies.

Anyone has any objection for this behaviour difference between
usage of ::regclass and to_regclass()?


With Regards,
Amit Kapila.
EnterpriseDB: http://www.enterprisedb.com