From: | "David E(dot) Wheeler" <david(at)kineticode(dot)com> |
---|---|
To: | Darren Duncan <darren(at)DarrenDuncan(dot)net> |
Cc: | PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: function_name.parameter_name |
Date: | 2010-09-09 16:12:12 |
Message-ID: | C55B942B-913E-4332-9FC6-3E952E37FE9D@kineticode.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Sep 8, 2010, at 3:57 PM, Darren Duncan wrote:
>> While I don't agree with the idea of providing extra names that are
>> probably mostly going to increase the confusion of someone trying to
>> understand such a system, I think this use case would be well covered by
>> synonyms. But these would be defined by a new SQL command, say CREATE
>> SYNONYM, not by funny notation on the initial CREATE FUNCTION call.
Sounds handy.
> Yes, and having a more general solution like CREATE SYNONYM is more important to have anyway. My "|" is simply a syntactic shorthand for a special case of CREATE SYNONYM, with respect to schema objects, and would parse into the same thing. I don't feel any need now for me to push this shorthand further. -- Darren Duncan
I can't get excited about it. I'm just happy the functionality is there. I was able to both simplify my PL/pgSQL code *and* make it much clearer what it's doing:
http://github.com/theory/pgxn-manager/commit/e136ccb342010e836c39dafa43b802478be445a0
That said, I'm assuming that the function-name block is really a RECORD object representing the argument signature. I could see a case for PL/pgSQL just having an "ARGS" variable or something that does the same thing. Kind of like triggers have NEW and OLD. But given that the functionality is already there, that's just gravy. Or sugar. I'm not sure which. Sugary gravy.
Best,
David
From | Date | Subject | |
---|---|---|---|
Next Message | Mark Llewellyn | 2010-09-09 16:22:51 | BUG #5650: Postgres service showing as stopped when in fact it is running |
Previous Message | Joe Conway | 2010-09-09 16:11:31 | Re: "serializable" in comments and names |