Re: Functions have 32 args limt ??? - Mailing list pgsql-general

From Richard Huxton
Subject Re: Functions have 32 args limt ???
Date
Msg-id 200308291140.35688.dev@archonet.com
Whole thread Raw
In response to Functions have 32 args limt ???  ("Ivar" <ivar@lumisoft.ee>)
List pgsql-general
On Friday 29 August 2003 11:12, Ivar wrote:
> > Well, a quick recompile and you're on your way again.
>
> You forgot that windows users aren't goot at compileing, ... .
> Windows user want running system.
> I'm sure because of it many windows users won't migrate to linux.

Many windows users shouldn't be administering a database. If recompiling a
package is unacceptable then you probably want to outsource your admin needs.
Open-source solutions aren't Microsoft solutions, and you need to think about
them differently.
The beauty of open-source is exactly that you have the source and can make
these changes. If you had a standard no-source package then you'd have to
wait for the next product release to get changes (if at all).

Now - if you installed from RPM and need some help compiling PG for the first
time then I'm sure we can help. In fact - if this is the case you can use the
source RPMs and build your own RPM package that you can install wherever you
like. You could even put it on a website so that anyone else who needs 100
parameter functions doesn't have to recompile.

> > * Personally I agree with those that think if your function has 32
> > parameters you need to rethink your function,
>
> You can see my function exaples, what can you suggest ?

I couldn't see what it did (I don't remember seeing any code, just
declarations). If it's just inserting a record, I can't see why you'd use a
function. It's probably a style thing - I tend to use functions for triggers
and very occasionally unusually complex select queries. Everything else goes
through SQL and is abstracted in the "middleware" layer of my apps.

> > Hmm - if it caught you out (*) then perhaps it needs to be clearly
> > documented.
>
> Best way isn't document this, but fix it to support more args - then it
> never be problem and doesn't need documenting.

Well - if you could come up with a way to do so without impacting other users,
I'm sure the developers would listen. Otherwise, it's going to be difficult
to persuade them to hurt 99% of users to help the 1% when there is already a
simple work-around.

--
  Richard Huxton
  Archonet Ltd

pgsql-general by date:

Previous
From: Martijn van Oosterhout
Date:
Subject: Re: process big
Next
From: Steven Tower
Date:
Subject: select taking forever