Horton

Re: FmgrInfo allocation patterns (and PL handling as staged programming)

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Chapman Flack <jcflack@acm.org>
Cc: PostgreSQL Hackers <pgsql-hackers@postgresql.org>
Date: 2025-04-06T19:47:46Z
Lists: pgsql-hackers 
Chapman Flack <jcflack@acm.org> writes:
> Can a spread-out variadic "any" arg list ever vary
> in length or type on the fly at a single call site?

Don't think so.

> I notice that funccache only hashes the first nargs
> argument types.

Yeah; the languages it serves today don't support VARIADIC ANY,
so it's not really a concern.  But if you'd like to consider doing so,
we could talk about what's needful.  I don't think funccache is set in
stone (yet).

			regards, tom lane

Commits

  1. Fix erroneous construction of functions' dependencies on transforms.

  2. Change SQL-language functions to use the plan cache.