On Thu, 2003-02-27 at 15:02, Tom Lane wrote:
> Rod Taylor <rbt@rbt.ca> writes:
> > Agreed -- but they're stored on the row themselves. You might be able
> > to write a function which executes dirty reads on the table and tells
> > you if the row is locked or not, but it's not going to be simple.
>
> Actually, I don't think you need a dirty read at all. A locked row
I see. That will make it quite a bit easier then. Perhaps I'll write a
function sometime. It would make it useful for fetching things out of a
persistent work queue. Right now I deal with userlocks -- but those can
be clumsy.
--
Rod Taylor <rbt@rbt.ca>
PGP Key: http://www.rbt.ca/rbtpub.asc