On Wed, Sep 24, 2008 at 07:33:27PM -0400, Tom Lane wrote:
> It will accept *definitions*, yes. I can't imagine why you'd think
> it's a good idea to not throw error when asked for an unknown variable's
> value.
because this is how it worked.
i'm not saying it was good. it worked that way, and introducing such
change made some code (rownum in this example) not working.
i understand that postgresql is more about correctness than end-user
niceness, which has some obvious benefits, but i think - in such cases
at least a one-line information in release docs wouldn't be out of
place.
> The whole custom-variable thing is being abused far beyond what the
> facility was intended for, anyway.
i think it's great idea. somebody thought about cool thing for modules,
other people found other uses for it.
right now they are being effectively punished for
creativeness (unfortunatelly i can't say it is me, as i got the variable
trick from somebody else).
best regards,
depesz
--
Linkedin: http://www.linkedin.com/in/depesz / blog: http://www.depesz.com/
jid/gtalk: depesz@depesz.com / aim:depeszhdl / skype:depesz_hdl / gg:6749007