Re: [HACKERS] NULL as an argument in plpgsql functions (fwd)

Oleg Bartunov <oleg@sai.msu.su> writes:
> Is this a bug or feature ? 

It's a bug.  Fixing it will require a wholesale code revision, however
(see my prior postings about redesigning the function call interface).

This is something we need to do for 6.6, IMHO, not only because of
the NULL-argument issue but also because it will solve the portability
problems that are being created by the existing fmgr interface (Alpha
bugs, need to dumb down to -O0 on some platforms, etc).  I've been
trying to summon the will to get started on it, but other things keep
getting in the way...

> Also, it seems there is a  limitation to a number of arguments.

Yes, 8.  I'm not planning to do anything about that in the near term.
Even just making the limit configurable would be a lot of work :-(
        regards, tom lane