On Thu, Jun 04, 2009 at 11:17:39AM -0400, Tom Lane wrote:
> Sam Mason <sam@samason.me.uk> writes:
> > I think that with 8.3 at least I'm going to carry on putting
> > IMMUTABLE on where I think it should be. Consider:
>
> there are corner cases where it's useful to have the function
> marked correctly rather than sloppily.
I think I originally noticed it when I'd written an "implies" operator
(i.e. (NOT $1) OR $2, as I don't trust myself to remember the logic)
and queries were running slowly until I noticed that a view was using
it deep inside itself somewhere. Seems like views would trigger this
behavior pretty reliably.
Doesn't seem like it needs fixing too urgently though.
--
Sam http://samason.me.uk/