On 2016-12-08 17:38:38 -0500, Tom Lane wrote:
> Andres Freund <andres@anarazel.de> writes:
> > I'm wondering if it's not time for $subject:
> > - V0 causes confusion / weird crashes when PG_FUNCTION_INFO_V1 was
> > forgotten
> > - They have us keep weird hacks around just for the sake of testing V0
> > - they actually cost performance, because we have to zero initialize Datums, even if
> > the corresponding isnull marker is set.
> > - they allow to call arbitrary functions pretty easily
>
> If by the first point you mean "assume V1 when no info function is found",
> I object to that. If you mean you want to require an info function, that
> might be OK.
I mean throwing an error. Silently assuming V1 seems like a horrible
idea to me. It doesn't seem unlikely that we want to introduce a new
call interface at some point given the runtime cost of the current one,
and that'd just bring back the current problem.
> The habit of zero-initializing Datums has got exactly nothing to do with
> V0 functions; it's about ensuring consistent results and avoiding
> heisenbugs from use of uninitialized memory. I do not think we should
> drop it.
Well, V0 functions don't have a real way to get information about NULL,
and we allow non-strict V0 functions, so?
Regards,
Andres